FMTOC_TK.indd Page i 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
CALCULUS
L A U R A TA A L MA N
James Madison University
P E T E R K OH N
James Madison University
W. H. Freeman and Company
New York
FMTOC_TK.indd Page ii 03/12/12 4:11 PM user-f502
Senior Publisher: Ruth Baruth
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Library of Congress Control Number: 2012947365
Executive Editor: Terri Ward
Marketing Manager: Steve Thomas
Complimentary Copy:
Market Development Manager: Steven Rigolosi
ISBN-13: 978-1-4641-2963-6
Developmental Editors: Leslie Lahr, Katrina Wilhelm
ISBN-10: 1-4641-2963-0
Senior Media Editor: Laura Judge
Associate Editor: Jorge Amaral
Student Edition Hardcover:
Editorial Assistant: Liam Ferguson
ISBN-13: 978-1-4292-4186-1
Photo Editor: Ted Szczepanski
ISBN-10: 1-4292-4186-1
Cover Photo Researcher: Elyse Rieder
Cover Designer: Vicki Tomaselli
Student Edition Paperback:
Text Designer: Marsha Cohen
ISBN-13: 978-1-4641-5108-8
Illustrations: Network Graphics
ISBN-10: 1-4641-5108-3
Illustration Coordinator: Bill Page
Production Coordinator: Susan Wein
Student Edition Loose-leaf:
Project Management and Composition: Aptara
ISBN-13: 978-1-4641-4005-1
Printing and Binding: RR Donnelley
ISBN-10: 1-4641-4005-7
© 2014 by W. H. Freeman and Company
All rights reserved
Printed in the United States of America
First printing
W. H. Freeman and Company
41 Madison Avenue
New York, NY 10010
Houndmills, Basingstoke RG21 6XS, England
www.whfreeman.com
FMTOC_TK.indd Page iii 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
D E D I C A T I O N
To Leibniz and Newton
—Laura Taalman
To Newton and Leibniz
—Peter Kohn
FMTOC_TK.indd Page iv 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
C O N T E N T S
Contents
iv
Features
About the Authors
vi
Acknowledgments
xiv
Preface
vii
To the Student
xvi
Media and Supplements
ix
*Starred titles indicate optional material
I DIFFERENTIAL CALCULUS
xi
Chapter 0
3.4
3.5
3.6
Functions and Precalculus* / 1
Chapter Review, Self-Test, and Capstones
0.1
0.2
0.3
0.4
0.5
Functions and Graphs
Operations, Transformations, and Inverses
Algebraic Functions
Exponential and Trigonometric Functions
Logic and Mathematical Thinking*
Chapter Review, Self-Test, and Capstones
2
19
35
47
63
73
Chapter 1
Limits / 77
1.1
1.2
1.3
1.4
1.5
1.6
An Intuitive Introduction to Limits
Formal Definition of Limit
Delta–Epsilon Proofs*
Continuity and Its Consequences
Limit Rules and Calculating Basic Limits
Infinite Limits and Indeterminate Forms
Chapter Review, Self-Test, and Capstones
78
90
100
109
123
138
152
Derivatives / 155
2.1
2.2
2.3
2.4
2.5
2.6
An Intuitive Introduction to Derivatives
Formal Definition of the Derivative
Rules for Calculating Basic Derivatives
The Chain Rule and Implicit Differentiation
Derivatives of Exponential and Logarithmic
Functions
Derivatives of Trigonometric and Hyperbolic
Functions
Chapter Review, Self-Test, and Capstones
156
169
187
201
3.1
3.2
3.3
iv
The Mean Value Theorem
The First Derivative and Curve Sketching
The Second Derivative and Curve Sketching
Definite Integrals / 315
4.1
4.2
4.3
4.4
4.5
4.6
4.7
Addition and Accumulation
Riemann Sums
Definite Integrals
Indefinite Integrals
The Fundamental Theorem of Calculus
Areas and Average Values
Functions Defined by Integrals
Chapter Review, Self-Test, and Capstones
402
Techniques of Integration / 407
5.4
5.5
5.6
5.7
Integration by Substitution
Integration by Parts
Partial Fractions and Other Algebraic
Techniques
Trigonometric Integrals
Trigonometric Substitution
Improper Integrals
Numerical Integration*
224
Chapter 6
236
Applications of Integration / 499
240
250
264
317
328
342
354
364
375
388
Chapter 5
Chapter Review, Self-Test, and Capstones
Applications of the Derivative / 239
313
Chapter 4
212
Chapter 3
278
291
302
II INTEGRAL CALCULUS
5.1
5.2
5.3
Chapter 2
Optimization
Related Rates
L’Hôpital’s Rule
6.1
6.2
6.3
6.4
6.5
Volumes by Slicing
Volumes by Shells
Arc Length and Surface Area
Real-World Applications of Integration
Differential Equations*
Chapter Review, Self-Test, and Capstones
408
420
432
444
454
467
480
494
500
514
526
542
559
573
FMTOC_TK.indd Page v 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
III SEQUENCES AND SERIES
Chapter 7
Sequences and Series / 577
7.1
7.2
7.3
7.4
7.5
7.6
7.7
Sequences
Limits of Sequences
Series
Introduction to Convergence Tests
Comparison Tests
The Ratio and Root Tests
Alternating Series
579
594
606
617
626
633
641
Chapter Review, Self-Test, and Capstones
655
Contents
v
11.3 Unit Tangent and Unit Normal Vectors
11.4 Arc Length Parametrizations and Curvature
11.5 Motion
874
881
891
Chapter Review, Self-Test, and Capstones
900
V MULTIVARIABLE CALCULUS
Chapter 12
Multivariable Functions / 903
12.1 Functions of Two and Three Variables
12.2 Open Sets, Closed Sets, Limits, and
Chapter 8
Power Series / 659
8.1
8.2
8.3
8.4
Power Series
Maclaurin Series and Taylor Series
Convergence of Power Series
Differentiating and Integrating Power Series
Chapter Review, Self-Test, and Capstones
660
672
681
694
12.3
12.4
12.5
12.6
12.7
904
Continuity
Partial Derivatives
919
933
Directional Derivatives and Differentiability
The Chain Rule and the Gradient
Extreme Values
Lagrange Multipliers
946
955
966
977
Chapter Review, Self-Test, and Capstones
987
703
Chapter 13
IV VECTOR CALCULUS
Double and Triple Integrals / 991
Chapter 9
Parametric Equations, Polar Coordinates,
and Conic Sections / 707
9.1
9.2
9.3
9.4
9.5
Parametric Equations
Polar Coordinates
Graphing Polar Equations
Computing Arc Length and Area with
Polar Functions
Conic Sections*
Chapter Review, Self-Test, and Capstones
708
724
733
749
758
13.1
13.2
13.3
13.4
13.5
13.6
Double Integrals over Rectangular Regions
Double Integrals over General Regions
Double Integrals using Polar Coordinates
Applications of Double Integrals
Triple Integrals
Integration using Cylindrical and Spherical
Coordinates
13.7 Jacobians and Change of Variables
992
1006
1017
1029
1041
Chapter Review, Self-Test, and Capstones
1081
1058
1069
774
Chapter 14
Chapter 10
Vector Analysis / 1085
Vectors / 777
10.1
10.2
10.3
10.4
10.5
10.6
Cartesian Coordinates
Vectors
Dot Product
Cross Product
Lines in Three-Dimensional Space
Planes
778
792
803
814
827
836
Chapter Review, Self-Test, and Capstones
847
14.1
14.2
14.3
14.4
14.5
14.6
Vector Fields
Line Integrals
Surfaces and Surface Integrals
Green’s Theorem
Stokes’ Theorem
The Divergence Theorem
1086
1097
1109
1122
1134
1143
Chapter Review, Self-Test, and Capstones
1152
Answers to Odd-Numbered Problems
Index
A-1
I-1
Chapter 11
Vector Functions / 851
11.1 Vector-valued Functions
11.2 The Calculus of Vector Functions
852
862
FMTOC_TK.indd Page vi 03/12/12 4:11 PM user-f502
A B O U T
T H E
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
A U T H O R S
Laura Taalman and Peter Kohn are professors of mathematics at James Madison University,
where they have taught calculus for a combined total of over 30 years.
Laura Taalman received her undergraduate degree from the University of Chicago
and master’s and Ph.D. degrees in mathematics from Duke University. Her research includes singular algebraic geometry, knot theory, and the mathematics of games and puzzles. She is a recipient of both the Alder Award and the Trevor Evans award from the
Mathematical Association of America, and the author of five books on Sudoku and the
mathematics of Sudoku. In her spare time, she enjoys being a geek.
Peter Kohn received his undergraduate degree from Antioch College, a master’s
degree from San Francisco State University, and a Ph.D. in mathematics from the University of Texas at Austin. His main areas of research are low-dimensional topology and knot
theory. He has been a national judge for MathCounts since 2001. In his spare time, he
enjoys hiking and riding his bicycle in the beautiful Shenandoah Valley.
vi
FMTOC_TK.indd Page vii 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
P R E F A C E
Calculus books have become full of clutter, distracting margin notes, and unneeded features. This calculus book clears out that clutter so that students can focus on the important
ideas of calculus. Our goal was to create a clean, streamlined calculus book that is accessible and readable for students while still upholding the standards required in science,
mathematics, and engineering programs, and that is flexible enough to accommodate different teaching and learning styles.
Linear Flow with Clean Margins
One thing that is distinctive about this calculus book is that it follows a linear writing style.
Figures and equations flow with the text as part of a clear, structured exposition instead of
being scattered about in the margins. We feel that this approach greatly increases the clarity of the book and encourages focused reading.
Exposition Before Calculation
Another distinctive feature of this book is that in each section we have separated the exposition and illustrative examples from the longer, more complicated calculational examples.
Including these longer examples separately from the exposition increases flexibility:
Students who want to read and understand the development of the material can do so
without being bogged down or distracted by large examples, while students who want to
use the book as a reference for looking up examples that are similar to homework problems can also do that.
Examples to Learn From
Within the exposition of each section are short examples that quickly illustrate the concepts
being developed. Following the exposition is a set of detailed, in-depth examples that explore
both calculations and concepts. We took great pains to provide many steps and illustrations
in each example in order to aid the student, including details about how to get started on a
problem and choose an appropriate solution method. One of the elements of the book that we
are most proud of is the “Checking the Answer” feature, which we have included after selected
examples to encourage students to learn how to check their own answers.
Building Mathematics
We were very careful in this book to approach mathematics as a discipline that is developed logically, theorem by theorem. Whenever possible, theorems are followed by proofs
that are written to be understood by students. We have included these proofs because
they are part of the logical development of the material, but we have clearly labeled and
indented each proof to indicate that it can be covered or skipped, according to instructor
preference. Each exercise set contains an optional subsection of proofs, many of which are
accessible even to beginning students. In addition, we have emphasized the interconnections among topics by providing “Thinking Back” and “Thinking Forward” exercises in
each section and “Capstone” problems at the end of each chapter.
Consistency and Reliability
Another improvement in this book is that it has a consistent and predictable structure. For
example, instructors can rely on every section concluding with a “Test Your Understanding”
vii
FMTOC_TK.indd Page viii 03/12/12 4:11 PM user-f502
viii
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Preface
feature which includes five questions that students can use to self-test and that instructors can choose to use as pre-class questions. The exercises are always consistently split
into subsections of different types of problems: “Thinking Back,” “Concepts,” “Skills,”
“Applications,” “Proofs,” and “Thinking Forward.” In addition, the “Concepts” subsection always begins with a summary exercise, eight true/false questions, and three example
construction exercises. Instructors and students alike can rely on this consistent structure
when assigning exercises and choosing a path of study.
Flexibility
We recognize that instructors use calculus books in many different ways and that the real
direction of a calculus course comes from the instructor, not any book. The streamlined,
consistent structure of this book makes it easy to use with a wide variety of courses and
pedagogical styles. In particular, instructors will find it easy to include or omit sections,
proofs, examples, and exercises consistently according to their preferences and course requirements. Students can focus on mathematical development or on examples and calculations as they need to throughout the course. Later, they can use the book as a reliable
reference.
We think it will be immediately clear to anyone opening this book that what we have
written is substantially different from the other calculus books on the market today while
still following the standard topics taught in most modern science, mathematics, and engineering calculus courses. Our hope is that faculty who use the book will find it flexible for
different pedagogical approaches and that students will be able to read it on different levels
as they learn to understand the beauty of calculus.
A Special Taalman/Kohn Option for Underprepared
Calculus Students
Do some of your calculus students struggle with algebra and precalculus
material? The Taalman/Kohn Calculus series has a ready-made option for such
students, called Calculus I with Integrated Precalculus. This option includes all
the material in Chapters 0–6 of Taalman/Kohn Calculus, but in a different order
and with supplementary precalculus and algebra material.
CALCULUS I
with Integrated Precalculus
L A U R A TA A L M A N
X Chapters 0–3 of Calculus I with Integrated Precalculus cover the same develop-
ment of differential calculus topics as Chapters 0–3 in Taalman/Kohn Calculus,
but the more complicated calculational examples are deferred to later chapters.
X Chapters 4–6 of Calculus I with Integrated Precalculus revisit differential calculus
through the lens of studying progressively more challenging types of functions.
Any exercises or examples from Taalman/Kohn Calculus that were left out of
Chapters 0–3 of Calculus I with Integrated Precalculus are included in Chapters 4–6. The requisite background precalculus and algebra material is built
from the ground up.
X Chapters 7–9 of Calculus I with Integrated Precalculus are identical to Chapters 4–6
of Taalman/Kohn Calculus and cover all topics from integral calculus.
Students who learn Calculus I from Calculus I with Integrated Precalculus can
continue with Calculus II using Taalman/Kohn Calculus or any other calculus textbook.
Students who have weak algebra and precalculus skills can succeed in STEM-level calculus if given the right help along the way, and Calculus I with Integrated Precalculus is
written specifically to address the needs of those students.
For an examination copy of Calculus I with Integrated Precalculus, please contact
your local W. H. Freeman & Company representative.
FMTOC_TK.indd Page ix 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
M E D I A
A N D
S U P P L E M E N T S
For Instructors
Instructor’s Solutions Manual
Single-variable ISBN: 1-4641-5017-6
Multivariable ISBN: 1-4641-5018-4
Contains worked-out solutions to all exercises in the text.
Test Bank
Computerized (CD-ROM), ISBN: 1-4641-2547-3
Includes multiple-choice and short-answer test items.
Instructor’s Resource Manual
ISBN: 1-4641-2545-7
Provides suggested class time, key points, lecture material, discussion topics, class activities, worksheets, and group projects corresponding to each section of the text.
Instructor’s Resource CD-ROM
ISBN: 1-4641-2548-1
Search and export all resources by key term or chapter. Includes text images, Instructor’s
Solutions Manual, Instructor’s Resource Manual, and Test Bank.
For Students
Student Solutions Manual
Single-variable ISBN: 1-4641-2538-4
Multivariable ISBN: 1-4641-5019-2
Contains worked-out solutions to all odd-numbered exercises in the text.
Software Manuals
Maple™ and Mathematica® software manuals are available within CalcPortal. Printed versions of these manuals are available through custom publishing. They serve as basic introductions to popular mathematical software options and guides for their use with Calculus.
Book Companion Web Site at www.whfreeman.com/tkcalculus
For students, this site serves as a FREE 24–7 electronic study guide, and it includes such
features as self-quizzes and interactive applets.
Online Homework Options
www.webassign.net/whfreeman
WebAssign Premium integrates the book’s exercises into the world’s most popular and
trusted online homework system, making it easy to assign algorithmically generated
homework and quizzes. Algorithmic exercises offer the instructor optional algorithmic
ix
FMTOC_TK.indd Page x 03/12/12 8:21 PM user-f502
x
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Media and Supplements
solutions. WebAssign Premium also offers access to resources, including the new Dynamic
Figures, CalcClips whiteboard videos, tutorials, and “Show My Work” feature. In addition,
WebAssign Premium is available with a fully customizable e-Book option that includes
links to interactive applets and projects.
www.yourcalcportal.com
CalcPortal combines a fully customizable e-Book, exceptional student and instructor resources, and a comprehensive online homework assignment center. Included are algorithmically generated exercises, as well as Precalculus diagnostic quizzes, Dynamic Figures,
interactive applets, CalcClips whiteboard videos, student solutions, online quizzes, Mathematica and Maple manuals, and homework management tools, all in one affordable, easyto-use, and fully customizable learning space.
webwork.maa.org
W. H. Freeman offers approximately 2,500 algorithmically generated questions (with full solutions) through this free, open-source online homework system at the University of Rochester. Adopters also have access to a shared national library test bank with thousands of
additional questions, including 1,500 problem sets matched to the book’s table of contents.
Additional Media
This easy-to-use Web-based version of the Instructor’s Solutions Manual allows instructors to generate a solution file for any set of homework exercises. Solutions can be downloaded in PDF format for convenient printing and posting.
Interactive e-Book at ebooks.bfwpub.com/tkcalculus
The Interactive e-Book integrates a complete and customizable online version of the text
with its media resources. Students can quickly search the text, and they can personalize the e-Book just as they would the print version, with highlighting, bookmarking, and
note-taking features. Instructors can add, hide, and reorder content, integrate their own
material, and highlight key text.
Course Management Systems
W. H. Freeman and Company provides courses for Blackboard, WebCT (Campus Edition
and Vista), Angel, Desire2Learn, Moodle, and Sakai course management systems. These
are completely integrated solutions that you can easily customize and adapt to meet your
teaching goals and course objectives. Visit www.macmillanhighered.com/catalog/other/
coursepack for more information.
This two-way radio frequency classroom response system was developed by educators
for educators. University of Illinois physicists Tim Stelzer, Gary Gladding, Mats Selen, and
Benny Brown created the i-clicker system after using competing classroom responses and
discovering that they were neither appropriate for the classroom nor friendly to the student. Each step of i-clicker’s development has been informed by teaching and learning.
i-clicker is superior to other systems from both a pedagogical and a technical standpoint.
To learn more about packaging i-clicker with this textbook, contact your local sales representative or visit www.iclicker.com.
FMTOC_TK.indd Page xi 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
F E A T U R E S
Each section opens with a list of the three main section topics. The list provides a focus
and highlights key concepts.
3.3
THE SECOND DERIVATIVE AND CURVE SKETCHING
Using first and second derivatives to define and detect concavity
The behavior of the first and second derivatives at inflection points
Using the second-derivative test to determine whether critical points are maxima, minima, or neither
Definitions are clearly boxed, numbered, and labeled for easy reference. To reinforce their
importance and meaning, definitions are followed by brief, often illustrated, examples.
DEFINITION 3.9
Formally Defining Concavity
Suppose f and f are both differentiable on an interval I.
(a) f is concave up on I if f is increasing on I.
(b) f is concave down on I if f is decreasing on I.
How does this formal definition of concavity correspond with our intuitive notion of concavity? Consider the functions graphed next. On each graph four slopes are illustrated and
estimated. Notice that when f is concave up, its slopes increase from left to right, and when
f is concave down, its slopes decrease from left to right.
Slopes increase when f is concave up
y
Slopes decrease when f is concave down
y
1 ⫺1
⫺3
4
⫺1
3
⫺4
1
x
x
Theorems are developed intuitively before they are stated formally, and simple examples
inform the discussion. Proofs follow most theorems, although they are optional, given
instructor preference.
THEOREM 3.4
Rolle’s Theorem
If f is continuous on [a, b] and differentiable on (a, b), and if f (a) = f (b) = 0, then there
exists at least one value c ∈ (a, b) for which f (c) = 0.
Actually, Rolle’s Theorem also holds in the more general case where f (a) and f (b) are equal
to each other (not necessarily both zero). For example, Rolle’s Theorem is also true if f (a) =
f (b) = 5, or if f (a) = f (b) = −3, and so on, because vertically shifting a function by adding
a constant term does not change its derivative. However, the classic way to state Rolle’s
Theorem is with f (a) and f (b) both equal to zero.
Proof. Rolle’s Theorem is an immediate consequence of the Extreme Value Theorem from
Section 1.4 and the fact that every extremum is a critical point. Suppose f is continuous on the
closed interval [a, b] and differentiable on the open interval (a, b), with f (a) = f (b) = 0. By the Extreme Value Theorem, we know that f attains both a maximum and a minimum value on [a, b]. If
one of these extreme values occurs at a point x = c in the interior (a, b) of the interval, then x = c
is a local extremum of f . By the previous theorem, this means that x = c is a critical point of f .
Since f is assumed to be differentiable at x = c, it follows that f (c) = 0 and we are done.
It remains to consider the special case where all of the maximum and minimum values of
f on [a, b] occur at the endpoints of the interval (i.e., at x = a or at x = b). In this case, since
f (a) = f (b) = 0, the maximum and minimum values of f (x) must both equal zero. For all x in [a, b]
we would have 0 ≤ f (x) ≤ 0, which means that f would have to be the constant function f (x) = 0
on [a, b]. Since the derivative of a constant function is always zero, in this special case we have
f (x) = 0 for all values of c in (a, b), and we are done.
xi
FMTOC_TK.indd Page xii 04/12/12 9:34 PM user-f502
xii
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Features
Color is used consistently and pedagogically in graphs and figures to relate like concepts.
For instance, the color used for rectangles in Riemann sum approximations is also quite
purposefully used for linear approximations of arc length and rectangular solid approximations of volume.
y
y
(f(a), g(a))
10
z
z
(f(t k⫺ 1), g(t k⫺ 1))
5
(f(tk ), g(tk ))
1
2
3
4
5
x
(f(b), g(b))
⫺5
x
y
x
x
y
Cautions are appropriately placed at points in the exposition where students typically
have questions about the nuances of mathematical thinking, processes, and notation.
CAUTION
1
x
It is important to note that although we use the notation x −1 to denote the reciprocal , the
notation f −1
does not stand for the reciprocal
1
f
of f . The notation f −1
used in Definition 0.10
is pronounced “f inverse.” We are now using the same notation for two very different
things, but it should be clear from the context which one we mean.
Every section includes short illustrative examples as part of the discussion and development of the material. Once the groundwork has been laid, more complex examples and
calculations are provided. Students find this approach easier to handle because the difficult
calculations do not interfere with the development of why things work. Example solutions
are explained in detail and include all the steps necessary for student comprehension.
EXAMPLE 4
Using critical points and Rolle’s Theorem to find local extrema
The function f (x) = x (x − 1)(x − 3) is a cubic polynomial with one local maximum and one
local minimum. Use Rolle’s Theorem to identify intervals on which these extrema exist.
Then use derivatives to find the exact locations of these extrema.
SOLUTION
The roots of f (x) = x (x − 1)(x − 3) are x = 0, x = 1, and x = 3. Since f is a polynomial,
it is continuous and differentiable everywhere. Therefore Rolle’s Theorem applies on the
intervals [0, 1] and [1, 3], and it tells us that at least one critical point must exist inside each
of these intervals.
The critical points of f are the possible locations of the local extrema that we seek. To
find the critical points we must solve the equation f (x) = 0. It is simpler to do some algebra
before differentiating:
f (x) =
d
d
(x (x − 1)(x − 3)) = (x 3 − 4x 2 + 3x) = 3x 2 − 8x + 3.
dx
dx
By the quadratic formula, we have f (x) = 0 at the points
x=
−(−8) ±
√
√
82 − 4(3)(3)
8 ± 28
4± 7
=
=
.
2(3)
6
3
These x-values are approximately x ≈ 0.451 and x ≈ 2.215. If we look at the graph of f ,
then we can see that the smaller of these two x-values is the location of the local maximum
and the larger is the location of the local minimum; see the figure that follows.
Following many example solutions, Checking the Answer encourages students to learn
to check their work, using technology such as a graphing calculator when appropriate.
FMTOC_TK.indd Page xiii 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Features
CHECKING
THE ANSWER
xiii
The graph of f (x) = x(x − 1)(x − 3) is shown next. Notice that the local extrema do seem
to occur at the values we just found.
Extrema at x ≈ 0.451 and x ≈ 2.215
4
4
⫺1
⫺8
Each section closes with five Test Your Understanding questions that test students on the
concepts and reading presented in the section. Because answers are not provided, instructors may choose to use these questions for discussion or assessment.
TEST YOUR
? UNDERSTANDING
Why could we not give a precise mathematical definition of concavity before this section?
The domain points x = c where f (c) = 0 or where f (c) does not exist are the critical
points of the function f . Why?
Why is it not clear to say a sentence such as “Because it is positive, it is concave up”?
How could this information be conveyed more precisely?
Why does it make sense that f is increasing when f is positive?
Suppose x = c is a critical point with f (c) = 0. Why does it make graphical sense that
f has a local minimum at x = c when f is concave up in a neighborhood around x = c?
Section Exercises are provided in a consistent format that offers the same types of exercises within each section. This approach allows instructors to tailor assignments to their
course, goals, and student audience.
Thinking Back exercises ask students to review relevant concepts from previous sections and lessons.
Concepts exercises are consistently formatted to start with the following three problems:
• Problem 0 tests understanding.
• Problem 1 consists of eight true/false questions.
• Problem 2 asks the student to create examples based on their understanding of the
reading.
Skills exercises offer ample practice, grouped into varying degrees of difficulty.
Applications exercises contain at least two in-depth real-world problems.
Proofs exercises can be completed by students in non-theoretical courses. Hints are
often provided, and many exercises mimic work presented in the reading and examples.
Often, these exercises are a continuation of a proof offered as a road map in the narrative.
Thinking Forward exercises plant seeds of concepts to come. In conjunction with the
Thinking Back exercises, they offer a “tie together” of both past and future topics, thereby providing a seamless flow of concepts.
Chapter Review, Self-Test, and Capstones, found at the end of each chapter, present
the following categories:
Definitions exercises prompt students to recall definitions and give an illustrative example.
Theorems exercises ask students to complete fill-in-the-blank theorem statements.
Formulas, Notation, and/or Rules exercises vary according to chapter content and
ask students to show a working understanding of important formulas, equations, notation, and rules.
Skill Certification exercises provide practice with basic computations from the chapter.
Capstone Problems pull together the essential ideas of the chapter in more challenging mathematical and application problems.
FMTOC_TK.indd Page xiv 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
A C K N O W L E D G M E N T S
There are many people whose contributions to this project have made it immeasurably
better. We are grateful to the many instructors from across the United States and Canada
who have offered comments that assisted in the development of this book:
Jabir Abdulrahman, Carleton University
Jay Abramson, Arizona State University
Robert F. Allen, University of Wisconsin–La Crosse
Roger Alperin, San Jose State University
Matthew Ando, University of Illinois
Jorge Balbas, California State University, Northridge
Lynda Ballou, New Mexico Institute of Mining and
Technology
E. N. Barron, Loyola University Chicago
Stavros Belbas, University of Alabama
Michael Berg, Loyola Marymount University
Geoffrey D. Birky, Georgetown University
Paul Blanchard, Boston University
Joseph E. Borzellino, California Polytechnic State University,
San Luis Obispo
Eddie Boyd, Jr., University of Maryland Eastern Shore
James Brawner, Armstrong Atlantic State University
Jennifer Bready, Mount Saint Mary College
Mark Brittenham, University of Nebraska
Jim Brown, Clemson University
John Burghduff, Lone Star College–CyFair
Christopher Butler, Case Western Reserve University
Katherine S. Byler Kelm, California State University,
Fresno
Weiming Cao, The University of Texas at San Antonio
Deb Carney, Colorado School of Mines
Lester Caudill, University of Richmond
Leonard Chastkofsky, The University of Georgia
Fengxin Chen, University of Texas at San Antonio
Dominic Clemence, North Carolina A&T State University
A. Coffman, Indiana–Purdue Fort Wayne
Nick Cogan, Florida State University
Daniel J. Curtin, Northern Kentucky University
Donatella Danielli-Garofalo, Purdue University
Shangrong Deng, Southern Polytechnic State University
Hamide Dogan-Dunlap, The University of Texas at El Paso
Alexander Engau, University of Colorado, Denver
Said Fariabi, San Antonio College
John C. Fay, Chaffey College
Tim Flaherty, Carnegie Mellon University
Stefanie Fitch, Missouri University of Science & Technology
Kseniya Fuhrman, Milwaukee School of Engineering
Robert Gardner, East Tennessee State University
Richard Green, University of Colorado, Boulder
Weiman Han, University of Iowa
Yuichi Handa, California State University, Chico
Liang (Jason) Hong, Bradley University
Steven Hughes, Alabama A&M University
Alexander Hulpke, Colorado State University
Colin Ingalls, University of New Brunswick, Fredericton
xiv
Lea Jenkins, Clemson University
Lenny Jones, Shippensburg University
Heather Jordan, Illinois State University
Mohammad Kazemi, The University of North Carolina
at Charlotte
Dan Kemp, South Dakota State University
Boris L. Kheyfets, Drexel University
Alexander A. Kiselev, University of Wisconsin–Madison
Greg Klein, Texas A&M University
Evangelos Kobotis, University of Illinois at Chicago
Alex Kolesnik, Ventura College
Amy Ksir, US Naval Academy
Dan Kucerovsky, University of New Brunswick
Trent C. Kull, Winthrop University
Alexander Kurganov, Tulane University
Jacqueline La Vie, SUNY College of Environmental Science
and Forestry
Melvin Lax, California State University, Long Beach
Dung Le, The University of Texas at San Antonio
Mary Margarita Legner, Riverside City College
Denise LeGrand, University of Arkansas–Little Rock
Mark L. Lewis, Kent State University
Xiezhang Li, Georgia Southern University
Antonio Mastroberardino, Penn State Erie, The Behrend
College
Michael McAsey, Bradley University
Jamie McGill, East Tennessee State University
Gina Moran, Milwaukee School of Engineering
Abdessamad Mortabit, Metropolitan State University
Emilia Moore, Wayland Baptist University
Vivek Narayanan, Rochester Institute of Technology
Rick Norwood, East Tennessee State University
Gregor Michal Olsavsky, Penn State Erie, The Behrend
College
Rosanna Pearlstein, Michigan State University
Kanishka Perera, Florida Institute of Technology
Cynthia Piez, University of Idaho
Jeffrey L. Poet, Missouri Western State University
Joseph P. Previte, Penn State Erie, The Behrend College
Jonathan Prewett, University of Wyoming
Elise Price, Tarrant County College
Stela Pudar-Hozo, Indiana University of Northwest
Don Redmond, Southern Illinois University
Dan Rinne, California State University, San Bernardino
Joe Rody, Arizona State University
John P. Roop, North Carolina A&T State University
Amber Rosin, California State Polytechnic University,
Pomona
Nataliia Rossokhata, Concordia University
Dev K. Roy, Florida International University
FMTOC_TK.indd Page xv 03/12/12 4:11 PM user-f502
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
Acknowledgments
Hassan Sedaghat, Virginia Commonwealth University
Asok Sen, Indiana University–Purdue University
Adam Sikora, The State University of New York at Buffalo
Mark A. Smith, Miami University
Shing Seung So, University of Central Missouri
David Stowell, Brigham Young University–Idaho
Jeff Stuart, Pacific Lutheran University
Howard Wainer, Wharton School of the University of
Pennsylvania
Thomas P. Wakefield, Youngstown State University
Bingwu Wang, Eastern Michigan University
Lianwen Wang, University of Central Missouri
Antony Ware, University of Calgary
xv
Talitha M. Washington, Howard University
Mary Wiest, Minnesota State University, Mankato
Mark E. Williams, University of Maryland Eastern Shore
G. Brock Williams, Texas Tech University
Dennis Wortman, University of Massachusetts, Boston
Hua Xu, Southern Polytechnic State University
Wen-Qing Xu, California State University, Long Beach
Yvonne Yaz, Milwaukee School of Engineering
Hong-Ming Yin, Washington State University
Mei-Qin Zhan, University of North Florida
Ruijun Zhao, Minnesota State University, Mankato
Yue Zhao, University of Central Florida
Jan Zijlstra, Middle Tennessee State University
We would also like to thank the Math Clubs at the following schools for their help in
checking the accuracy of the exercises and their solutions:
CUNY Bronx Community College
Duquesne University
Fitchburg State College
Florida International University
Idaho State University
Jackson State University
Lander University
San Jose State University
Southern Connecticut State University
Texas A&M University
Texas State University–San Marcos
University of North Texas
University of South Carolina–Columbia
University of South Florida
University of Wisconsin–River Falls
Our students and colleagues at James Madison University have used preliminary versions of this text for the past two years and have helped to clarify the exposition and remove
ambiguities. We would particularly like to thank our colleagues Chuck Cunningham, Rebecca Field, Bill Ingham, John Johnson, Brant Jones, Stephen Lucas, John Marafino, Kane
Nashimoto, Edwin O’Shea, Ed Parker, Gary Peterson, Katie Quertermous, James Sochacki,
Roger Thelwell, Leonard Van Wyk, Debra Warne, and Paul Warne for class-testing our book
and for their helpful feedback. During the class-testing at JMU, hundreds of students provided feedback, made suggestions that improved the book, and, of course, showed us how
they learned from the book! Thank you to all of our students, especially Lane O’Brien and
Melissa Moxie for their meticulous review of an earlier draft of the text.
Chris Brazfield, now at Carroll Community College, helped with the initial development of the text and the ideas behind it. Kevin Cooper of Washington State University contributed many interesting and challenging real-world applications, and Elizabeth
Brown and Dave Pruett of James Madison University contributed greatly to the development of the chapter on vector calculus. Roger Lipsett of Brandeis University wrote the
excellent solution manual for the text, and at the same time eliminated any ambiguities in
the exercises. We owe all of them great thanks for their expertise.
We also owe thanks to all of the people at W. H. Freeman who helped with the development of this text. Our developmental editor, Leslie Lahr, has been with this project from
the beginning. Even under pressure, Leslie always maintains a positive attitude and finds a
way for us to move forward. Without her support, we would not have made it through the
rocky patches. Our executive editor, Terri Ward and developmental editor Katrina Wilhelm
helped keep us on track while we wrote, rewrote, revised, revised, and revised some more,
and we thank them for their support and patience. Brian Baker, our meticulous copy editor, made significant improvements to the text. Misplaced commas, dangling modifiers,
and run-on sentences didn’t stand a chance under his scrutiny. Ron Weickart and his team
at Network Graphics took our graphs and sketches and turned them into the beautiful
artwork in the text. Sherrill Redd and the compositing team at Aptara did a great job
implementing all the design elements from our crazy LaTeX files.
Finally, we would like to thank our families and friends for putting up with us during
the years of stress, turmoil, and tedium that inevitably come with any book project. Without their support, this book would not have been possible.
FMTOC_TK.indd Page xvi 03/12/12 4:11 PM user-f502
T O
T H E
/203/MH01832/mor34949_disk1of1/0078034949/mor34949_pagefiles
S T U D E N T
Learning something new can be both exciting and daunting. To gain a full understanding
of the material in this text, you will have to read, you will have to think about the connections between the new topics and the topics that were previously presented, and you will
have to work problems—many, many problems.
The structure of this text should help you understand the material. The material is
laid out in a linear fashion that we think will facilitate your understanding. Each section
is separated into two main parts: first, a presentation of new material and then second, a
set of Examples and Explorations, where you will find problems that are carefully worked
through. Working through these examples on your own, as you read the steps for guidance,
will help prepare you for the exercises.
Reading a mathematics book isn’t like reading a novel: You may have to read some
parts more than once, and you may need to make notes or work things out on paper. Pay
special attention to the “Checking Your Answer” features, so that you can learn how to
check your own answers to many types of questions.
To succeed in calculus, you need to do homework exercises. The exercises in every
section of this text are broken into six categories: “Thinking Back,” “Concepts,” “Skills,”
“Applications,” “Proofs,” and “Thinking Forward.”
• As the title suggests, the Thinking Back problems are intended to tie the current material
to material you’ve seen in previous sections or even previous courses.
• The Concepts problems are designed to help you understand the main ideas presented
in the section without a lot of calculation. Every group of Concepts exercises begins by
asking you to summarize the section, continues with eight true/false questions, and
then asks for three examples illustrating ideas from the section.
• The bulk of the exercises in each section consists of Skills problems that may require
more calculation.
• The Applications exercises use the concepts from the section in “real-world” problems.
• The Proofs exercises ask you to prove some basic theory from the section.
• Finally, the Thinking Forward questions use current ideas to introduce topics that you
will see in subsequent sections.
We hope this structure allows you to tie together the material as you work through
the book. We have supplied the answers to the odd-numbered exercises, but don’t restrict
yourself to those problems. You can check answers to even-numbered questions by hand
or by using a calculator or an online tool such as wolframalpha.com. After all, on a quiz or
test you won’t have the answers, so you’ll have to know how to decide for yourself whether
or not your answers are reasonable.
Some students may like to work through each section “backwards,” starting by
attempting the exercises, then checking back to the examples as needed when they get
stuck, and, finally, using the exposition as a reference when they want to see the big picture.
That is fine; although we recommend that you at least try reading through the sections in
order to see how things work for you. Either way, we hope that the separation of examples
from exposition and the division of homework problems into subsections will help make
the process of learning this beautiful subject easier. We have written this text with you, the
student, in mind. We hope you enjoy using it!
xvi
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
C H A P T E R 0
Functions and Precalculus
0.1
Functions and Graphs
4
(2, 4)
What Is a Function?
Vertical and Horizontal Line Tests
Properties of Graphs
Examples and Explorations
0.2
2
Operations, Transformations, and Inverses
Combinations of Functions
Transformations and Symmetry
Inverse Functions
Examples and Explorations
0.3
Algebraic Functions
Power Functions
Polynomial Functions
Rational Functions
Absolute Value Functions
Examples and Explorations
0.4
Exponential and Trigonometric Functions
Exponential Functions
Logarithmic Functions
Trigonometric Functions
Inverse Trigonometric Functions
Examples and Explorations
0.5
Logic and Mathematical Thinking*
From Definitions to Theorems
Quantifiers
Implications
Counterexamples
Simple Mathematical Proofs
Examples and Explorations
A =⇒ B
Chapter Review, Self-Test, and Capstones
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
2
Chapter 0
0.1
FUNCTIONS AND GRAPHS
November 26, 2012
Functions and Precalculus
Definition of functions and their domain and range
Graphs, horizontal and vertical line tests, and one-to-one-functions
Graphical properties and features, asymptotes, and average rate of change
What Is a Function?
Mathematics is a language. In order to understand it, you have to learn how to read it and
speak it with the correct vocabulary. Since calculus is at its heart the study of functions
of real numbers, the universe we will spend most of our time exploring is the set of real
numbers and the relationships between sets of real numbers. Therefore we must begin by
setting out the mathematical language that describes these relationships we call “functions.” Once we all speak the same language, we can start building the theory of calculus.
Functions and their properties will be at the core of everything we study in this text.
In previous courses you likely encountered functions that were given in terms of formulas,
such as
y(x) = x 2 ,
that relate two variables x and y. To set the stage for studying such functions, we must
first be more precise about what functions are. Instead of thinking of functions merely as
formulas, think of them as describing a certain kind of rule, relationship, or mapping from
the elements of one set to the elements of another set.
DEFINITION 0.1
Functions
A function f from a set A to a set B is an assignment f that associates to each element
x of the domain set A exactly one element f (x) of the codomain, or target, set B.
We will use the notation
f:A→B
to represent a function f together with its domain set A and target set B. This notation is
pronounced “f from A to B.” If x and y are variables that represent elements of the sets A
and B, respectively, then we say that y is a function of x and write y = f (x) or y(x).
The variable x is called the independent variable and represents the “input” of the
function. The function f sends each input x to one and only one “output,” some value of
the dependent variable y. Notice that y depends on x, according to the assignment defined
by the function f .
For example, the assignment f : R → R that squares real numbers is a function, since
each real number x is assigned to one and only one real-number square x 2 . Here R denotes the set of all real numbers, and f assigns each real-number input to exactly one
real-number output. Some real numbers (such as 3 and −3) get sent to the same square
( f (−3) = f (3) = 9), but this does not violate the definition of function. You can think of a
function as a machine that takes any given input value x and produces exactly one output
value f (x) (pronounced “f of x”), shown as follows:
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
x
Functions and Graphs
3
f (x)
3
x2
f
3
9
x2
9
So, what isn’t a function? If a rule assigns a real-number input to more than one
√ output, then that rule is not a function. For example, consider the formula y = ± x. This
assignment does not define y as a function of x, because the input x = 4 corresponds to
two different y-values, both y = −2 and y = 2. In the “function machine” type of illustration just shown, the number 4 would go into the machine and two numbers, −2 and 2,
would come out at once as outputs. This situation is not allowed for functions.
Returning to the squaring function y = x 2 , notice that some real numbers can never
serve as outputs, because squares of real numbers can never be negative. The range, or set
of possible outputs, of the squaring function is [0, ∞). In this text we will usually work with
functions whose domains and ranges are unspecified subsets of real numbers and whose
rules are given by formulas such as f (x) = x 2 .
DEFINITION 0.2
Domain and Range of a Function
If f is a function between unspecified subsets of R, then we will take the domain of f to
be the largest subset of R for which f is defined:
Domain( f ) = { x ∈ R | f (x) is defined }.
The range of such a function is the set of all possible outputs that it can attain:
Range( f ) = { y ∈ R | there is some x ∈ Domain( f ) for which f (x) = y }.
√
For example, the function f (x) = x − 1 is defined only when x ≥ 1, and therefore f (x) =
√
x − 1 has domain [1, ∞). When we write the square root symbol without
the “±” before
√
it, we always mean the positive square root. This means that f (x) = x − 1 can attain only
√
nonnegative y-values. In fact, every nonnegative value √
can be expressed in the form x − 1
for some value of x, and therefore the function f (x) = x − 1 has range [0, ∞).
A few notes about the notation we just used: The curly-brackets notation used in Definition 0.2 is called set notation, and it is a way to describe a set of real numbers. In this
case the set notation for the domain of f is pronounced “the set of all x contained in R such
that f (x) is defined.” Notice in particular that the symbol “∈” means “contained in” and
the vertical bar means “such that.”
TECHNICAL POINT The name of a function is usually a single letter, such as “f .” The name
of the output of a function f evaluated at an input x is “f (x).” In this situation f is a function,
or relationship, and f (x) is a number that represents the output of the function at the input
value x. However, it is sometimes convenient to write f (x) (the name of the output of the
function) instead of f (the name of the function itself). This allows us to indicate the name
we are using for the independent variable when we reference the function. We may also
write things like “consider the function f (x) = x 2 + 1,” by which we mean “consider the
function f whose output at a real number x is f (x) = x 2 + 1.”
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
4
Chapter 0
November 26, 2012
Functions and Precalculus
Vertical and Horizontal Line Tests
A function whose domain and range are sets of real numbers can be represented as
a collection of pairs (x, f (x)) of real numbers. If we plot these pairs as points in the
Cartesian plane, we obtain the graph of the function. In general we have the following
definition:
DEFINITION 0.3
The Graph of a Function
The graph of a function f is the collection of ordered pairs (x, f (x)) for which x is in the
domain of f . In set notation we can write
Graph( f ) = { (x, f (x)) | x ∈ Domain( f ) }.
For example, the graph of f (x) = x 2 is the collection of ordered pairs of the form (x, x 2 ), for
x ∈ R. Since f (−1) = (−1)2 = 1 and f (2) = 22 = 4, the points (−1, 1) and (2, 4) are on the
graph of f (x) = x 2 . In contrast, the point (1, 2) is not a part of the graph, because f (1) = 2,
as shown in the following graph:
Graph of f (x) = x 2 and partial table of values
y
(2, 4)
4
x
3
2
(1, 1)
2
1
1
1
2
(1, 2)
1
x2
1
1 (not 2)
4
(1, 1)
1
2
x
A function always has exactly one output value for every input in the domain, which
means that the graph of a function always passes the following test, which you will prove in
Exercise 90:
THEOREM 0.4
The Vertical Line Test
A graph represents a function if and only if every vertical line intersects the graph in at
most one point.
For example, consider the three graphs that follow this paragraph. The leftmost graph
passes the vertical line test and thus is the graph of a function. The graph in the middle fails the vertical line test because the vertical line x = 2 intersects the graph in two
points, (2, 1) and (2, 3); therefore the middle graph does not represent a function. The rightmost graph assigns the same output to two distinct inputs, but that is perfectly fine for a
function. Because the graph on the right passes the vertical line test, it is the graph of a
function.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
December 1, 2012
0.1
A graph that is a function
A graph that is not a function
y
A function, but not one-to-one
y
y
4
4
4
3
3
3
2
2
2
1
1
1
1
2
3
4
x
5
Functions and Graphs
1
2
3
4
x
1
2
3
4
x
If each element of the range of a function is the output of exactly one element of the
domain, then the function is said to be one-to-one. Graphically, we can tell if a function f is
one-to-one by checking to see if it passes the horizontal line test: if f is one-to-one, then
every horizontal line meets the graph of f at most once; see Exercise 91. Algebraically, this
means that a function f is one-to-one if two distinct elements in the domain are always
sent to different elements of the range:
DEFINITION 0.5
One-to-One Function
A function f is one-to-one if, for all a and b in the domain of f ,
a = b =⇒ f (a) = f (b).
In this definition the notation ⇒ is pronounced “implies,” and it means that if the
left-hand part of the expression is true, then the right-hand part of the expression is also
true. In other words, the statement “A ⇒ B” is synonymous with the statement “if A,
then B.”
A logically equivalent form of Definition 0.5 is its so-called contrapositive:
f (a) = f (b) =⇒ a = b.
As we will see in Section 0.5, the contrapositive of an implication A ⇒ B is the equivalent
statement (not B) ⇒ (not A). The contrapositive form of Definition 0.5 is often easier to
use, because it is an affirmative rather than a negative statement. For example, f (x) = 3x
is one-to-one because if 3a = 3b, then we can guarantee that a = b. In contrast, the squaring function f (x) = x 2 is not one-to-one, because we cannot guarantee that if a2 = b2 , then
a = b (for example, (−3)2 = 32 , but −3 = 3).
Properties of Graphs
The table that follows gives us vocabulary and precise mathematical definitions for various
types of graphical behavior. Rows 1, 2, 5, and 6 describe behaviors that a function could
exhibit at a specific point. The remaining rows describe graphical behaviors that occur over
an interval I of real numbers. Much of the material in the early chapters of this book will
be dedicated to developing techniques for properly defining and identifying these types of
CONFIRMING PAGES
12:3
TKmaster2010
WHF00153/FREE087-Taalman
6
Chapter 0
November 26, 2012
Functions and Precalculus
properties of functions. For now we present them just to set terminology and to familiarize
ourselves with various types of function behavior.
Vocabulary
Definition
Behavior
f has a root at x = c
f (c) = 0
graph intersects the
x-axis at x = c
f has a y-intercept
at y = b
f (0) = b
graph intersects the
y-axis at y = b
f is positive on I
f (x) > 0 for all x ∈ I
graph is above the x-axis on I
f is increasing on I
f (b) > f (a)
for all b > a in I
graph moves up as we
look from left to right on I
f has a local
maximum at x = c
f (c) ≥ f (x)
for all x near x = c
graph has a relative “hilltop”
at x = c
f has a global
maximum at x = c
f (c) ≥ f (x)
for all x ∈ Domain( f )
graph is the highest
at x = c
f is concave up on I
will state precisely in
Section 3.3
graph curves upwards on I
like part of a “U”
f has an inflection
point at x = c
will state precisely in
Section 3.3
graph of f changes
concavity at x = c
Of course, there are similar definitions for local and global minima and for negative, decreasing, and concave-down behavior; see Exercises 20 and 21. Notice that we describe
extrema (maxima and minima) by where on the x-axis they occur, since we can always find
the corresponding y-values from these x-values. The concept of “near” in the description
of a local maximum will be made more precise in Chapters 1 and 2. Inflection points and
concavity cannot be precisely defined until we learn about derivatives in Chapters 2 and 3.
In that chapter we will also learn ways for algebraically calculating the locations of extrema
and inflection points. Until then, we will have to be content with examining such things
graphically.
For example, the list that follows at the right describes some aspects of the graphical
behavior of the graph y = f (x) shown on the left.
y
4
3
2
1
4 3 2 1
1
2
3
1
2
3
4
x
roots at x = −3, x ≈ −0.4, and x = 3
y-intercept at y = −1
local maxima at x = −2 and x = 3
global maximum at x = −2
inflection points at x = −1 and x = 2
positive on (−3, −0.4)
increasing on (−∞, −2) and (1, 3)
concave up on (−1, 2)
In fact, technically the function f graphed at the left is increasing on the larger intervals
(−∞, −2] and [1, 3]. This is because we do have f (b) > f (a) for all values b > a in these
closed intervals. Most of the time we will be concerned only with the open intervals on which
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
Functions and Graphs
7
a function is increasing or decreasing, but there will be a few times in later chapters when
we need to consider closed or half-closed intervals. For now, we will use open intervals
unless we require otherwise.
The increasing and/or decreasing behavior of a function is related to its average rate of
change on various intervals. The average rate of change of a function f on an interval [a, b]
measures how much the output f (x) changes over that interval. Average rates of change
will be extremely important in Chapter 2 when we study the derivative.
DEFINITION 0.6
Average Rate of Change
The average rate of change of a function f on an interval [a, b] is the slope of the line
from (a, f (a)) to (b, f (b)), which is given by the quotient
f (b) − f (a)
.
b−a
For example, the function whose properties we just enumerated is increasing on the interval
(1, 3), moving up from (1, f (1)) = (1, −2) to (3, f (3)) = (3, 0). The average rate of change tells
us how much the function increased per unit change in the input, on average:
f (3) − f (1)
0 − (−2)
=
=1
3−1
2
unit up for every unit across. We can also measure average rate of change over intervals
where the function both increases and decreases; for example, with the same function, on
the interval [−3, 3] there is an average rate of change of
f (3) − f (−3)
0−0
=
=0
3 − (−3)
6
units up for every unit across; look at the graph to see why this makes sense.
Sometimes a graph gets closer and closer to a horizontal or vertical line, or asymptote.
In Chapter 1, we will define asymptotes precisely, using limits. For now, we will use the
following definition: A line l is an asymptote of a function f if the difference between the
graph of l and the graph of f gets as small as we want as either x or y increases in magnitude.
For example, the following graph of a function f has vertical asymptotes at x = −2 and
x = 2, and a horizontal asymptote at y = 1:
A function with three asymptotes
y
5
4
3
2
1
5 4 3 2 1
1
1
2
3
4
5
x
2
3
Notice that a graph can cross one of its horizontal asymptotes; the preceding graph above
does so at the point (0, 1). This is just one of the reasons that we are avoiding using the
overly loose definition of asymptote that you may have heard in previous courses (“an
asymptote is a line that the graph gets infinitely close to, but never reaches”).
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
8
Chapter 0
November 26, 2012
Functions and Precalculus
Examples and Explorations
EXAMPLE 1
Identifying functions and their domains and ranges
Determine whether or not each of the following relationships is a function. For each relationship that is a function, describe its natural domain and range and determine whether
or not it is one-to-one. For each relationship that is not a function, describe the parts of
the definition of a function that are violated.
(a) The rule g : R → R that assigns each real number x to the numbers whose square is x.
(b) The relationship defined by this table:
x
P(x)
1
5
2
2
3
9
4
−1
5
0
6
9
(c) Let P be the set of all living people in the world, and let W be the set of all women
that have ever lived. Define f : P → W so that each person is assigned to his or her
biological mother.
√
(d) f (x) = 2 − x + 5
(e) h(x) =
1
x2 − 4
SOLUTION
(a) This rule is not a function, for two reasons. First of all, negative real numbers do not
have real square roots, so g is not defined on the given domain√of R. Second,
each
√
positive number x has two numbers whose square is x, namely, x and − x, so this
rule would not send each domain element to exactly one output.
(b) The relationship P(x) defined by the table is a function, because the table assigns each
value in the domain {1, 2, 3, 4, 5, 6} to exactly one element of the range {−1, 0, 2, 5, 9}.
This function is not one-to-one because P(3) and P(6) are both equal to 9.
(c) This relationship is a function because each person has one exactly one woman who
is his or her biological mother. No person is without a biological mother, and no
person has more than one biological mother. Here the domain is P and the range is
the subset of W consisting of women that have had biological children. This function
is not one-to-one, since there are examples of different people that have the same
biological mother.
(d) This rule is a function because for each value√x for which the formula makes sense, there
is exactly one real number described by 2 − x + 5. For x to be in the domain, we must
have x +5 ≥ 0 (since x +5 is under a square-root sign), and thus we must have x ≥ −5.
Therefore the domain of f is√
[−5, ∞). The range of y = f (x) is the set of y-values that
can occur as outputs.
Since
x + 5 can take on any value greater than or equal to 0,
√
the expression 2 − x + 5 can take on any value less than or equal to 2. Therefore the
range of f is (−∞, 2]. This function is one-to-one because if f (a) = f (b), then
√
√
2 − a + 5 = 2 − b + 5 =⇒
a+5= b+5
=⇒ a + 5 = b + 5
=⇒ a = b.
1
(e) The rule h(x) is a function because for each value x at which 2
is defined, there is
x −4
exactly one real number that h describes. The domain of h(x) is the set of all x-values
for which x 2 − 4 = 0 (since x 2 − 4 is in a denominator). Therefore the domain of
h is everything except x = ±2. To find the range of h(x) we must find the y-values
that can be expressed
in the form y = h(x) for some x. Solving for x in terms of y
we obtain x =
y = 0 and
1
y
1
y
+ 4. This means we can find an x that maps via f to y as long as
+ 4 ≥ 0. It can be shown that the solution of the latter inequality is
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
1
4
−∞, −
9
Functions and Graphs
1
∪ [0, ∞). Therefore the range of h(x) is −∞, − ∪ (0, ∞). This function
4
1
3
is not one-to-one because, for example, h(1) and h(−1) are both equal to − .
CHECKING
THE ANSWER
The following functions f and h have domains marked in blue on the x-axis and ranges
marked in red on the y-axis:
h hasdomain x =
±2
1
and range −∞, − ∪ (0, ∞)
f has domain [−5, ∞)
and range (−∞, 2]
4
y
10
EXAMPLE 2
y
2
1
1
0.5
5
5
10
x
4 3 2 1
1
1
0.5
2
1
2
3
4
x
Evaluating function notation
√
Given that f (x) = x 3 − x, evaluate f (2), f (a), f (x + 1), and f ( f (x)).
SOLUTION
√
To evaluate f (x) = x 3 − x at a given input, simply replace x in the formula with whatever
the input is:
√
√
f (2) = 2 3 − 2 = 2 1 = 2 ;
√
f (a) = a 3 − a ;
f (x + 1) = (x + 1) 3 − (x + 1) ;
√
√
f ( f (x)) = f (x) 3 − f (x) = (x 3 − x ) 3 − x 3 − x .
EXAMPLE 3
Finding a “good” graphing window
Use a graphing utility to find a graphing window that accurately represents the key features
of the graph of the function f (x) = x 3 − 6x 2 − x + 6.
SOLUTION
The three graphs that follow show y = f (x) in various graphing windows. Each of these
windows is “bad” in the sense that the true behavior of the graph of f is not represented.
f (x) = x 3 − 6x 2 − x + 6
x ∈ [−3, 3], y ∈ [−10, 10]
f (x) = x 3 − 6x 2 − x + 6
x ∈ [−100, 100], y ∈ [−50, 50]
10
⫺3
50
3
⫺10
f (x) = x 3 − 6x 2 − x + 6
x ∈ [−20, 20], y ∈ [−1000, 1000]
⫺100
1000
100
⫺20
⫺50
CONFIRMING PAGES
20
⫺1000
18:20
TKmaster2010
WHF00153/FREE087-Taalman
10
Chapter 0
November 26, 2012
Functions and Precalculus
A “good” window (if one exists) is a window in which the local behavior of the graph of f is
clear and the global behavior is accurately represented (the “ends” of the graph keep going
in the direction indicated). The following figure shows the graph of f in a “good” window:
f (x) = x 3 − 6x 2 − x + 6
x ∈ [−3, 7], y ∈ [−40, 40]
40
⫺3
7
⫺40
For now we will use trial and error to find an effective graphing window. We will be able to
be more systematic after we have learned more about derivatives and function behavior. EXAMPLE 4
Function behavior at points and on intervals
(a) Sketch the graph of a function that has the following characteristics:
roots at x = −2, x = 1, and x = 3
local minimum at x = −1
horizontal asymptote at y = −2
local maximum at x = 2
(b) Approximate the locations of the inflection points on your graph.
(c) Given the graph that you sketched, find the intervals on which f is
positive
increasing
decreasing
concave up
SOLUTION
(a) By plotting the points (−2, 0), (1, 0), and (3, 0), drawing a dashed asymptote at y = −2,
and plotting some low point for the function at x = −1 and some high point at x = 2
(the information in the problem does not tell us exactly how high or how low), one
might make the following sketch:
One possible graph of f
y
3
2
1
3 2 1
1
1
2
3
4
5
6
7
x
2
3
4
(b) The inflection points on this graph occur where the concavity of the graph changes
from a ∪ shape to a ∩ shape, or vice versa. These are the locations where the graph
1
“flexes.” In our graph, these points occur at approximately x = and x = 4. Note that
2
these points are also the locations of the steepest parts of the graph.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
0.1
11
Functions and Graphs
(c) Given our graph, the function f is positive on (−∞, −2) and (1, 3), increasing
on
1
(−1, 2), decreasing on (−∞, −1) and (2, ∞), and concave up on −∞,
and (4, ∞).
2
EXAMPLE 5
Finding the domain of a function by using equalities and inequalities
1 − 2x
Find the domain of the function f (x) =
.
x+1
SOLUTION
To find the domain of f , we ask which values of x can be validly plugged into the equation
that defines f (x). In order for the value of f to be defined for a real number x, that value of x
must make the quotient underneath the square-root sign nonnegative and the denominator nonzero. Thus the domain of f (x) is the set of real numbers that simultaneously satisfy
the following:
1 − 2x
≥0
x+1
and
x + 1 = 0.
1
The only x-values at which the quotient in the inequality can change sign are x = and
2
x = −1, since those are the values that make either the numerator or denominator equal
to zero. To determine the intervals on which the quotient is positive or negative we
1
need only check its sign between the possible change points x = and x = −1. For
example, evaluating the expression at x = −2, x = 0, and x = 1 gives
2
1 − 2(−2)
pos
=
= negative,
−2 + 1
neg
1 − 2(0)
pos
=
= positive,
0+1
pos
1 − 2(1)
neg
=
= negative.
1+1
pos
We can record this information on a number line with a sign chart as follows:
⫺
⫹
⫺1
⫺
1
2
Since the quotient in question is negative on (−∞, −1) and
defined on those intervals. Note that x =
1
2
1
,∞
2
, the function f (x) is not
is in the domain of f because there is no problem
taking the square root of zero. However, since we cannot divide by zero, the function
is not
defined at x = −1. Therefore the domain of f is the half-open interval −1,
EXAMPLE 6
1
2
.
A review of factoring techniques
Find the solution sets of each of the following equations:
(a) 2x 3 − 5x 2 − 3x = 0
(b) 3x 2 = 7x − 1
(c) 2x 5 − 32x = 0
SOLUTION
(a) The number of real number solutions of a polynomial equation is at most the value of
the degree, or highest power, of the polynomial. Therefore we can expect the equation
2x 3 − 5x 2 − 3x = 0 to have at most three real number solutions. The right-hand side
of this equation can be easily factored as:
2x 3 − 5x 2 − 3x = x(2x 2 − 5x − 3) = x(2x + 1)(x − 3).
CONFIRMING PAGES
11:1
TKmaster2010
WHF00153/FREE087-Taalman
12
Chapter 0
December 6, 2012
Functions and Precalculus
The expression 2x 3 − 5x 2 − 3x is zero if and only if one of x, 2x + 1, or x − 3 is zero. In
1
other words, the solutions of the equation 2x 3 − 5x 2 − 3x = 0 are x = 0, x = − , and
2
x = 3.
(b) We first need to write the equation 3x 2 = 7x−1 in the general form of a quadratic equation: 3x 2 − 7x + 1 = 0. This equation cannot be easily factored with the reverse “FOIL”
(first-outside-inside-last) method, so we’ll apply the quadratic formula, which says
that the roots of a quadratic equation of the form ax 2 + bx + c = 0 are
√
−b ± b2 − 4ac
x=
.
2a
In this example we have a = 3, b = −7, and c = 1, so the solutions of 3x 2 − 7x + 1 = 0
are
√
√
−(−7) ± (−7)2 − 4(3)(1)
7 ± 49 − 12
7 ± 37
x=
=
=
.
2(3)
6
6
√
√
1
1
Therefore, the solutions of 3x 2 − 7x + 1 = 0 are x = (7 + 37 ) and x = (7 − 37 ).
6
6
Clearly we could not have easily figured that out by doing the “FOIL” method
backwards!
(c) This time the factoring will involve two applications of the well-known factoring formula a2 − b2 = (a + b)(a − b) for the difference of two squares:
2x 5 − 32x = 0
2x(x 4 − 16) = 0
2x(x 2 − 4)(x 2 + 4) = 0
2x(x − 2)(x + 2)(x 2 + 4) = 0
← formula for a2 − b2 with a = x 2 and b = 4
← formula for a2 − b2 with a = x and b = 2
Thus 2x 5 − 32x = 0 whenever 2x = 0, x − 2 = 0, x + 2 = 0, or x 2 + 4 = 0. Note that
x 2 + 4 = 0 has no real solutions, because there is no real number that satisfies
x 2 = −4. Therefore the real-number solution set of the original equation 2x 5 −32x = 0
is {−2, 0, 2}.
CHECKING
THE ANSWER
To check the answers in Example 6, simply substitute each proposed solution into the
original equation. Each solution should satisfy the equation. For example, to check that
x = −2, x = 0, and x = 2 are solutions in part (c) of the example we note that
2(−2)5 − 32(−2) = 0,
← evaluate equation at x = −2
5
← evaluate equation at x = 0
5
← evaluate equation at x = 2
2(0) − 32(0) = 0,
2(2) − 32(2) = 0.
Of course, this will not tell you whether you have missed any solutions, but it will tell you
whether the solutions you found are correct.
EXAMPLE 7
Finding the average rate of change of a function on an interval
Calculate the average rate of change of the function f (x) = x 2 − x + 1 (a) on the interval
[0, 3] and (b) on the interval [−1, 1]. Then (c) illustrate these average rates of change
graphically.
SOLUTION
(a) Using the formula from Definition 0.6 we find that the average rate of change of f on
[0, 3] is
f (3) − f (0)
(32 − 3 + 1) − (02 − 0 + 1)
7−1
6
=
=
= = 2.
3−0
3−0
3−0
3
CONFIRMING PAGES
10:57
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
13
Functions and Graphs
(b) Using the same formula again, we find that the average rate of change of f on [−1, 1] is
f (1) − f (−1)
(12 − 1 + 1) − ((−1)2 − (−1) + 1)
1−3
−2
=
=
=
= −1.
1 − (−1)
1 − (−1)
1 − (−1)
2
Notice that the average rate of change of f (x) = x 2 − x + 1 is different, depending on
what interval we consider.
(c) Graphically, the two average rates we found can be represented as slopes of line
segments, as follows:
Slope of line from
(0, f (0)) to (3, f (3)) is 2
Slope of line from
(−1, f (−1)) to (1, f (1)) is −1
y
y
7
7
3
3
1
1
EXAMPLE 8
1
3
1
x
1
1
3
x
A function that is defined in pieces
A piecewise-defined function is a function that is defined in pieces, with different formulas
on different parts of its domain. Let f be the function defined piecewise by
2
x , if x ≤ −1
f (x) =
2x, if x > −1.
Find f (−5), f (−1), and f (3), and then sketch a graph of y = f (x).
SOLUTION
Since −5 ≤ −1, we have f (−5) = (−5)2 = 25. Since −1 ≤ −1, we have f (−1) = (−1)2 = 1.
In contrast, since 3 > −1, we have f (3) = 2(3) = 6. To graph f , we begin by graphing the
functions y = x 2 and y = 2x that are used in the definition of f , as shown in the first two
figures that follow:
y = 2x
y = x2
y
y = f (x)
y
y
3
3
3
2
2
2
1
3 2 1
1
1
1
1
2
3
x
3 2 1
1
1
2
3
x
3 2 1
1
2
2
2
3
3
3
1
2
3
x
To graph f , we must restrict the graph of y = x 2 to the interval (−∞, −1] and restrict
y = 2x to (−1, ∞). Whether these intervals are open or closed is important. To find f (−1) we
use the first equation y = x 2 : f (−1) = (−1)2 = 1. Note that when we sketched the graph
of f in the third figure, we used open and closed dots to represent the function values
corresponding to the ends of open and closed intervals in the domain, respectively.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
14
Chapter 0
EXAMPLE 9
November 26, 2012
Functions and Precalculus
Functions involving pairs or triples of real numbers*
In this example we examine functions with multivariable inputs and/or outputs. Let R2 be
the set of ordered pairs of real numbers, and R3 the set of ordered triples of real numbers.
We will not need to work with such functions until much later in this book, but we present
them here to illustrate the variety of functions we will be encountering in this course.
(a) Explain why the rule p : R → R3 defined by p(t) = (3t, t 2 − 1, t) is a function. Then
find p(2).
(b) Explain why the rule q : R2 → R2 defined by q(x, y) = (x − y, −3x) is a function. Then
find q(3, 2).
SOLUTION
(a) The rule p(t) is a function because every real number t gets sent to exactly one triple of
numbers. For example, p(2) = (3(2), 22 − 1, 2) = (6, 3, 2).
(b) Similarly, the rule q(x, y) is a function because each pair of numbers (x, y) in the domain
R2 gets sent to exactly one pair of numbers in the range R2 . For example, q(3, 2) =
(3 − 2, −3(3)) = (1, −9).
TEST YOUR
? UNDERSTANDING
Why might the notation f (x) be wrongly confused with the notation for a product?
How is it different from the notation for a product? What is the difference between the
notation f and the notation f (x)?
If a rule f assigns both 4 and 8 to the same output value, can that rule be a function?
Why or why not?
The vertical line test states that each vertical line has to intersect the graph in at most
one point. Is it okay for a vertical line to pass through the graph at no points, that is, for
the line not to intersect the graph?
How do we evaluate a function that is defined in pieces? That is, given a function f (x)
defined piecewise, how do we go about finding, say, f (2)?
Define each of the following with mathematical notation: function, one-to-one, global
maximum, asymptote, root.
EXERCISES 0.1
Thinking Back
Interval notation: Describe each of the following subsets of the
real numbers in interval notation.
x = ±3
x > −2 and x = 5
x < 0 or x ≥ 10
x < 3 and x = 4
Solving equations and inequalities: Find the solution sets of the
equations and inequalities that follow. Write your answers in
interval notation (or, if the solution is a discrete set of points,
a list of those points).
x
=0
x−2
x 3 − 5x 2 + 6x < 0
x
>0
x−2
x 3 − 5x 2 + 6x = 0
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: Functions are the same as equations.
(b) True or False: The domain of every function is a subset
of R.
(c) True or False: The function that for each x has output
f (x) = 1 is a one-to-one function.
(d) True or False: Every global maximum of a function is
also a local maximum.
(e) True or False: Every local minimum of a function is also
a global minimum.
(f) True or False: The graph of a function can never cross
one of its asymptotes.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
(g) True or False: Average rates of change can be thought
of as slopes.
(h) True or False: A function can have different average
rates of change on different intervals.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
3.
4.
5.
6.
7.
8.
(a) A function that is defined with a formula.
(b) A function that is not defined with a formula.
(c) A formula that does not define a function.
State the mathematical definition of a function, and
describe its meaning in your own words. Support your
answer with an example of something that is a function
and an example of something that is not.
Suppose P is the set of people alive today and C is the set
of possible eye colors. Let f : P → C be the rule that assigns to each person his or her eye color. Is f a function?
Why or why not?
Use set notation to define the domain of a function. Then
use the same
√ notation to express the domain of the function f (x) = x.
Use set notation to define the range of a function. Then
use the same notation to express the range of the function
f (x) = x 2 .
Determine whether the points (a) (3,
√ 2), (b) (1, 1), and (c)
(−5, 2) lie on the graph of f (x) = x + 1, without referring to a picture of the graph of f .
Describe the graph of the function f (x) = 3x + 2 as a set
of ordered pairs.
9. Consider the function f (x) = x 2 + 1.
(a) Explain why y = 5 is in the range of f .
(b) Explain why y = 0 is not in the range of f .
(c) Argue that the range of f (x) = x 2 + 1 is [1, ∞).
10. Determine whether or not each diagram that follows represents a function. If it does, find its domain and range,
and determine whether it is one-to-one. If it does not,
explain what goes wrong.
f
g
h
1
5
1
5
1
5
2
6
2
6
2
6
3
7
3
7
3
7
4
8
4
8
4
8
11. Construct a rule f : {2, 4, 6, 8, 10} → {1, 2, 3, 4} that is a
function. Express this function three ways: as a list, as a
table, and as a diagram. Is your function one-to-one?
What is its range?
12. Construct a rule f : {2, 4, 6, 8, 10} → {1, 2, 3, 4} that is not
a function. Justify your answer.
13. If the graph of a rule y = f (x) passes through (−2, 1) and
(2, 1), could that rule be a function? Why or why not?
14. A constant function is a function f : A → B with the
property that there is some b ∈ B for which f (x) = b for all
x ∈ A. (The output of the function is constantly the same.)
Describe
(a) a constant function f : R → R
(b) a constant function g : R → [−5, −2]
(c) a constant function h : R → R2
Functions and Graphs
15
15. The identity function for a set A is the function f : A → A
defined by f (x) = x (so called because the output is identical to the input). For which of the following domains and
ranges is there a well-defined identity function? Why or
why not?
(a) f : R → R
(b) g : R2 → R
(c) h : R3 → R3
16. Let P be the set of all people living in the United States.
Give examples of each of the following functions and
state their ranges:
(a)
(b)
(c)
(d)
(e)
the identity function f : P → P
two different constant functions g : P → P
a non-constant, non-identity function g : P → P
a constant function h : P → R
a non-constant function h : P → R
17. Explain in your own words why the vertical line test
determines whether a graph is a function.
18. Explain in your own words why the horizontal line test
determines whether a function is one-to-one.
19. Show that f (x) = x 2 + 1 is not one-to-one, using values
of f (not the horizontal line test).
20. Define what it means for a function f with domain R
to have (a) a global minimum at x = c and (b) a local minimum at x = c.
21. Define what it means for a function f with domain R to
be (a) negative on an interval I and (b) decreasing on an
interval I.
22. Make a labeled graph that illustrates why it makes sense
that a function is increasing on an interval I if, for all b > a
in I, we have f (b) > f (a). Include labels for a, b, f (a), and
f (b), and for the interval I.
23. How is the formula for average rate of change related to
the formula for computing slope?
24. Illustrate on a graph of f (x) = 1 − x 2 that the average rate
of change of f on [−1, 3] is −2.
25. For each local maximum x = c in the following graph,
approximate the largest possible δ > 0 so that f (c) ≥ f (x)
for all x ∈ (c − δ, c + δ). Similarly, for the one local minimum x = b, find the largest δ so that f (b) ≤ f (x) for all
x ∈ (b − δ, b + δ).
y
19
3
2
1
1
2
3
x
8
13
26. Use the definition of a local maximum to explicitly argue
why the function graphed in Exercise 25 does not have a
local maximum at x = 0.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
16
Chapter 0
November 26, 2012
Functions and Precalculus
Skills
Find the domain and range of each function in Exercises
27–32. Use a graphing utility or plot points to sketch a graph
of the function, and illustrate the domain and range on the
graph.
√
√
28. f (x) = x − 1
27. f (x) = x − 1
1
29. f (x) =
x+2
1
31. f (x) = 2
x +1
1
30. f (x) = √
5−x
1
32. f (x) = 2
x −1
Find the domain of each function in Exercises 33–42.
3x + 1
33. f (x) = x(x − 2)
34. f (x) =
2x − 1
1
35. f (x) = (x − 1)(x + 3)
36. f (x) = √
x2 − 4
1
1
37. f (x) = −4
38. f (x) =
(x − 1)(x + 3)
x2
√
x 3/4
x2 − 1
40. f (x) =
39. f (x) = 2
3x − 5
x −9
√
√
x2 − 1
x
1
42. f (x) = √
−
41. f (x) = √
x−2
x−1
x2 − 9
Use a graphing utility to sketch a graph of each function in Exercises 51–56. Use trial and error to find a graphing window so
that your graph represents the local and global behavior of the
function. Include the x and y ranges of your window in your
answer.
51. f (x) = x 2 − 0.1
52. f (x) = (x 2 − 5)7
53. f (x) = x − 3x − 7x
54. f (x) = x 3 − 11x 2 + 10x
55. f (x) = x 2 − 17x − 18
56. f (x) =
5
4
2x 2 − 2
x 2 − 3x − 5
Describe the key properties of each graph in Exercises 57–62,
including the following:
domain and range;
locations of roots, intercepts, local and global maxima
and minima, and inflection points;
intervals on which the function is positive or negative,
increasing or decreasing, and concave up or down;
any horizontal or vertical asymptotes.
y
57.
Evaluate each function in Exercises 43–47 at the values
indicated. Simplify your answers if possible. (Note: We will not
need to work with such functions until much later in the book, but
we present them here to illustrate the variety of functions we will
be encountering in this course.)
y
58.
3
3
2
2
1
1
1
1
2
x
3
1
1
2
3
4
x
1
43. If f (x) = x 2 + 1, find
(b) f (a3 )
(a) f (−4)
44. If k(x) =
(c) f ( f (x))
y
59.
x
, find
x +1
(a) k(5)
45. If l(a, b, c) =
√
(a) l(5, 3, 2)
(b) k(x + h)
2
2
(c) l(x, y, z)
(b) g(1)
(a) F(2, 3, 5)
(b) F(5, 2, 3)
1
1
2
3
2 1
1
x
y
61.
2
3
4
1
2
3
x
y
62.
(c) F(a, b, 0)
For each piecewise-defined function in Exercises 48–50,
(a) calculate f (−1), f (0), f (1), and f (2), and (b) sketch a graph
of f .
2
x + 3, if x < 0
48. f (x) =
3 − x, if x ≥ 0
⎧
⎨ 3x + 1, if x ≤ 0
4, if 0 < x ≤ 1
49. f (x) =
⎩
x 3 , if x > 1
⎧
⎨ 4x − 1, if x < 0
2, if x = 0
50. f (x) =
⎩
−3x + 5, if x > 0
1
3
4.75
(c) g(x + 1)
47. If F(u, v, w) = (3u + v, u − w, v + 2w), find
1
2
46. If g(v) = (v − 1, v, v2 ), find
(a) g(0)
3
3.25
2
(c) k(k(x))
a2 + b2 + c2 , find:
(b) l(3, 0, 4)
y
60.
7
2
3
2
2
1
1
3 2 1
1
x
2
4
4
x
3
Sketch the graph of functions f that satisfy the lists of conditions given in Exercises 63–72, if possible.
63. Domain R, concave up everywhere, and decreasing
everywhere.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.1
64. Domain R, concave down everywhere, and decreasing
everywhere.
65. Domain R, concave up everywhere, increasing everywhere, and negative everywhere.
66. Domain R, concave down everywhere, increasing everywhere, and negative everywhere.
67. Always increasing, with two horizontal asymptotes, one
at y = −2 and one at y = 2.
68. Domain (0, ∞), always negative, and always increasing.
69. Four roots but no y-intercept.
70. Concave down on (−∞, 2), concave up on (2, ∞), and
always increasing.
71. Concave down on (−∞, 0) and concave up on (0, ∞) but
without an inflection point at x = 0.
17
Functions and Graphs
72. Average rate of change of 3 on [0, 2], average rate of
change of −1 on [0, 1], and average rate of change of 0
on [−2, 2].
In Exercises 73–78, find the average rate of change of the function f on the interval [a, b].
73. f (x) = −0.5 + 4.2x, [a, b] = [1, 3.5]
74. f (x) = 3, [a, b] = [−100, 100]
√
75. f (x) = x + 1, [a, b] = [1, 9]
76. f (x) =
1−x
, [a, b] = [0, 0.5]
1 + x3
1
, [a, b] = [0.9, 1.1]
x
3
78. f (x) = (x − 2)2 + , [a, b] = [−2, 2]
x
77. f (x) =
Applications
In Exercises 79–82, sketch and label the graph of a function
that describes the given situation.
79. An island warthog population initially grows quickly, but
as space and food become sparse on the island, the population growth slows down. Eventually the population of
the island levels off at 512 warthogs.
80. Susie is late for calculus class and leaves her dorm in a panic.
She hurries towards the math building, but about halfway
there, she realizes she has left her notebook in her room. She
sprints back to her dorm and gets her notebook. Coming out
of the dorm, she sprains her ankle, so the best she can do is
limp as fast as she can to her classroom.
81. Suppose that after you drink a cup of coffee the amount of
caffeine in your body rises sharply and then decreases by
half every hour. You have one cup of coffee in the morning
and then no more.
82. On a dare, you go skydiving. Gravity causes you to fall
faster and faster as you plummet towards the ground.
When you open your parachute, your speed is drastically
reduced. After opening your parachute you approach the
ground at a constant speed.
(a) Graph your distance from the ground as a function of
time.
(b) Graph your velocity as a function of time.
83. For each situation described, identify any independent
and dependent variables, and express their relationship
as an equation in multivariable function notation (see
Example 9):
(a) H is the length of the hypotenuse of the right triangle
with legs of length a and b.
(b) V is the volume of a rectangular prism (“box”) with
dimensions x, y, and z.
84. If your rain-catching bucket starts empty and collects
3 inches of rain during a 6-hour rainstorm, what is the
average rate of change of the level of rainwater in the
bucket over the 6 hours that it rained? Did the rain necessarily collect in the bucket at a constant rate?
85. A disgruntled pet store owner abandoned an unknown
number of groundhogs on a small island in 1996. Since
then it has been determined that the average rate of
change of the groundhog population was 4 groundhogs
per year and that the groundhog population was a linear function of time. When the abandoned groundhogs
were discovered in 2001, there were 376 groundhogs on
the island. How many groundhogs did the disgruntled
pet store owner originally leave on the island?
86. The number N of operating drive-in movie theatres in the
state of Virginia in various years y is given in the table
below.
y 1958 1967 1972 1977 1982 1999
N
143
90
102
87
56
9
(a) Find the average rate of change in the number of
drive-in movie theatres in Virginia over each time interval between table entries.
(b) Describe the units and real-world significance of
these average rates of change.
(c) Over which time period was the average rate of
change the most drastic? On average, assuming that
no new theatres were built, how many drive-ins
closed per year during that period?
87. In your first job after graduating from college you make
$36,000 a year before taxes. After four years you get a raise
of $2,500. Two years after that you change jobs and go to
work for a company that pays you $49,000 a year.
(a) Construct a piecewise-defined function that describes your pretax income in the year that is t years
after you graduate from college.
(b) Write down a function that describes the total
amount of money you will have earned t years after
graduating from college.
(c) How many years after graduating from college will
you have earned a total of one million pre-tax dollars?
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
18
Chapter 0
November 26, 2012
Functions and Precalculus
88. The following table shows the year 2000 Federal Tax Rate
Schedule for single filers:
Taxable income:
Over
The federal tax owed is:
Not over Amount Plus % Of amt. over
$0 $26,250
$0
15%
$0
$26,250 $63,550
$3,937
28%
$26,250
$63,550 $132,600 $14,381
31%
$63,550
$132,600 $288,350 $35,787
36%
$132,600
$91,857 39.6%
$288,350
$288,350
——
(a) How much tax would you owe if you made $18,000
of taxable income? What if you made $180,000?
(b) What percentage of your taxable income did you owe
in taxes if your taxable income was $18,000? What if
your taxable income was $180,000?
(c) Construct a piecewise-defined function describing
the dollar amount of tax T owed by a single person
with m dollars of taxable yearly income. Each piece
of your function will be linear. Do the pieces “match
up”? Does this make financial sense?
Proofs
89. Use Definition 0.2 to prove that the range of the function
f (x) = 3x − 1 is R.
90. Use Definition 0.1 to prove that a graph represents a function if and only if it passes the vertical line test.
91. Use Definition 0.5 to prove that a function is one-to-one
if and only if its graph passes the horizontal line test.
92. Use the contrapositive form of Definition 0.5 to prove that
the function f (x) = 3x + 1 is one-to-one.
93. Use the definition of decreasing to prove that the function
f (x) = 1 − 3x is decreasing on (−∞, ∞).
94. Use the definition of increasing to prove that the function
f (x) =
1
is increasing on (−∞, 3).
3−x
95. Prove that the average rate of change of the linear function f (x) = −2x + 4 on any interval I is always equal
to −2.
96. Show that the average rate of change of every linear function f (x) = mx + b is constant, that is, the same over any
choice of interval. (Hint: Use [c, d] to denote the interval,
since the letter b is already used in the equation for f (x).)
Thinking Forward
Evaluations for slopes and derivatives: Evaluate each function at
the values indicated. Simplify your answers if possible.
If f (x) = 4 − x 2 , find
f (1 + 0.1) − f (1)
0.1
√
If f (x) = x, find
(a)
(a)
f (1 + h) − f (1)
h
(b)
f (1 + 0.001) − f (1)
0.001
y
(b)
f (x) − f (1)
x−1
12
(b) q(3, h)
8
4
(c) q(x, 0.5)
Evaluations for series: Evaluate each function at the values indicated. Simplify your answers if possible.
1
1
,5
2
If c(x, n) = 1 −
(b) S
1
,n
2
(c) S(x, 5)
x 2n
x4
x6
x2
+
−
+ · · · + (−1)n
,
2!
4!
6!
(2n)!
find
(a) c(π , 3)
(b) c(π , n)
(c) c(x, 3)
2
3
4
x
Find the equation of the tangent line for f (x) = 3x + 1
at x = 2. (Hint: Think about the graph.) Why is this not
surprising?
Find the equation of the tangent line for f (x) = 4 − x 2
at x = 0. Again, think about the shape of the graph
before you attempt to answer this question.
Use a graph to visually estimate the slope of the
tangent line for f (x) = x 2 at x = 2. Use this slope estimate to write down an approximate formula for the
tangent line to f (x) = x 2 at x = 2.
xn
x2 x3 x4
+
−
+ · · · + (−1)n+1 , find
If S(x, n) = x −
2
3
4
n
(a) S
16
(x + h)2 − x 2
, find
If q(x, h) =
h
(a) q(3, 0.5)
Tangent lines: The tangent line to the graph of a function f at
x = c is the line that passes through the point (c, f (c)) and has
slope determined by the “direction” that the graph is moving.
If you imagine a graph of y = f (x) as a hilly curve that a small
car is driving on, then the tangent line is the line determined
by the car’s headlights at time x = c. For example, the graph
that follows shows the tangent line for f (x) = x 2 at x = 2.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
0.2
Operations, Transformations, and Inverses
19
OPERATIONS, TRANSFORMATIONS, AND INVERSES
Constant multiples, sums, products, quotients, and compositions of functions
Translations, stretches, compressions, and reflections of graphs
Inverse functions and their properties
Combinations of Functions
So far we have been thinking of functions as operators, that is, a function f takes an input
x and operates on it to produce an output f (x). We now want to think of functions in a
different way, as objects that can be added to, subtracted from, and multiplied or divided
by one another. Of course, sums, differences, products, and quotients of functions f and g
will be defined in terms of sums, differences, products, and quotients of their outputs f (x)
and g(x).
For example, given two functions f and g, we can add them together and get a new
function f + g. How this new function operates on inputs will depend on how the original
two functions operated. In other words, to define the function f + g, we must say what
f + g does to each input x. If x is in the domain of f and in the domain of g, the obvious
choice is to define ( f + g)(x) to be the sum of f (x) and g(x). For example, if f (x) = x 2
and g(x) = 3x + 1, then for all values of x we define ( f + g)(x) to be f (x) + g(x) = x 2 +
3x + 1. In particular, this means that ( f + g)(2) is equal to the sum of f (2) = 22 = 4 and
g(2) = 3(2) + 1 = 7, so that ( f + g)(2) = 4 + 7 = 11. The other arithmetic operations work
similarly on functions:
DEFINITION 0.7
Arithmetic Combinations of Functions
Suppose f and g are functions and k is a real number.
(a) The constant multiple of f by k is the function kf defined by (kf )(x) = k f (x) for
all x in the domain of f .
(b) The sum of f and g is the function f + g defined by ( f + g)(x) = f (x) + g(x) for all
x in the domains of both f and g.
(c) The product of f and g is the function f · g defined by ( f · g)(x) = f (x)g(x) for
all x in the domains of both f and g.
f
f
f (x)
(d) The quotient of f and g is the function defined by
(x) =
for all x in the
domains of both f and g with g(x) = 0.
g
g
g(x)
There is an additional operation on functions that we do not have for numbers, called
composition. We compose two functions f and g by taking the output from one function
as the input for the other:
DEFINITION 0.8
The Composition of Two Functions
The composition of two functions f and g is the function f ◦ g defined by
( f ◦ g)(x) = f ( g(x))
for all x in the domain of g such that g(x) is in the domain of f .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
20
Chapter 0
November 26, 2012
Functions and Precalculus
For example, if f (4) = 6 and g(10) = 4, then ( f ◦ g)(10) = f (g(10)) = f (4) = 6. You should
think of compositions as nestings of functions. The notation ( f ◦ g) is pronounced
“f composed with g” or sometimes “f circle g.” The notation f ( g(x)) is pronounced
“f of g of x.”
If g : X → Y and f : Y → Z, then their composition is a function ( f ◦ g) : X → Z
that takes an input x first to g(x) and then to f ( g(x)). For example, if f (x) = x 2 and g(x) =
3x + 1 then
( f ◦ g)(x) = f ( g(x)) = f (3x + 1) = (3x + 1)2 .
Notice that although the function f appears first (i.e., on the left) in the notation, it is the
function g that gets applied to the input x first. Composition is not a commutative operation, which means that f ◦ g is not necessarily the same function as g ◦ f . With the same
example of f (x) = x 2 and g(x) = 3x + 1, if we compose in the other order, we get
( g ◦ f )(x) = g( f (x)) = g(x 2 ) = 3(x 2 ) + 1.
CAUTION
You may have noticed that the notation for composition looks a bit like multiplication
notation, but there is a key difference. When we want to denote multiplication we will
use a small closed dot or no dot at all. To denote composition of functions we will always
use an open circle.
Transformations and Symmetry
Another way we can obtain new functions from old is through transformations. Given
a function f (x) and constants C and k, we could consider such modifications as f (x) + C,
f (x + C), k f (x), and f (kx). Each of these transformations changes f (x) graphically and algebraically.
For example, transforming f (x) to f (x) ± C clearly adds C units to every output of f (x).
This means that the graph of y = f (x) shifts up or down vertically C units everywhere, to
become the graph of y = f (x) + C or y = f (x) − C, as illustrated by the red and green
graphs, respectively, shown in the figure next at the left. If we instead add a constant to the
independent variable and transform f (x) to f (x±C), the graph shifts left or right horizontally
by C units, as illustrated in the green and red graphs shown in the figure at the right. (Note
that the shift to the left for f (x + 2) and to the right for f (x − 2) might be the opposite of
what we might initially expect.) These additive transformations are called translations.
f (x) + 2 shifts up 2
f (x) − 2 shifts down 2
f (x + 5) shifts left 5
f (x − 5) shifts right 5
y
y
2
2
1
5
5
10
5
2
1
2
5
2
2
x
5
5
CONFIRMING PAGES
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
21
Operations, Transformations, and Inverses
If we instead transform f (x) by multiplication to k f (x), then the graph of y = f (x)
expands or contracts vertically by a factor of k to become the graph of y = k f (x), as shown
in the red and green graphs next at the left. In contrast, if we do the same transformation to
the independent variable and transform f (x) to f (k x), this contracts or expands the graph
of y = f (x) by a factor of k in the horizontal direction, as illustrated in the red and green
graphs next at the right.
f (2x) compresses horizontally by 2
1 f x stretches horizontally by 2
2
2f (x) stretches vertically by 2
1
f (x) compresses vertically by 2
2
y
y
2
2
2
6
2
4
2
2
2
2
2
4
1
2
1
1
2
x
6
4
2
2
6
x
2
1
4
2
6
2
What happens if we multiply x or f (x) by a negative number? We can answer that question by just looking at what happens when we multiply by −1. Changing f (x) to −f (x)
transforms all positive outputs into negative outputs, and vice versa. The graph of y = f (x)
is then reflected across the x-axis to become the graph of y = −f (x), as shown in the red
graph in the figure that follows. If we instead multiply the independent variable by −1,
then we obtain a reflection across the y-axis, as shown in the green graph.
−f (x) reflects across the x-axis
f (−x) reflects across the y-axis
y
6
4
2
6
4
2
2
4
6
x
2
4
6
Now if we want to transform f (x) to f (−2x), for example, we can transform f (x) first to f (2x)
and then by reflection to f (−2x).
The table that follows summarizes the graphical and algebraic effects of the transformations just discussed. You will prove these general results in Exercises 86–88.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
22
Chapter 0
November 26, 2012
Functions and Precalculus
Transformation
Graphical Result
Algebraic Result
f (x) + C
shifts up C units if C > 0
shifts down C units if C < 0
(x, y) → (x, y + C)
f (x + C)
shifts left C units if C > 0
shifts right C units if C < 0
(x, y) → (x − C, y)
k f (x)
vertical stretch by k if k > 1
vertical compression by k if 0 < k < 1
(x, y) → (x, ky)
f (k x)
horizontal compression by k if k > 1
horizontal stretch by k if 0 < k < 1
(x, y) →
−f (x)
graph reflects across the x-axis
(x, y) → (x, −y)
f (−x)
graph reflects across the y-axis
(x, y) → (−x, y)
1
x, y
k
Some graphs do not change under certain transformations. For example, the graph of
f (x) = x 2 shown next at the left remains the same if we reflect it across the y-axis. We say
that this function has y-axis symmetry. As another example, the graph of g(x) = x 3 shown
at the right remains the same if we reflect it first across the y-axis and then across the x-axis.
f (x) = x 2 preserved under y-axis reflection
g(x) = x 3 preserved under 180◦ rotation
y
y
3
3
2
2
1
1
3 2 1
1
1
2
3
x
3 2 1
1
2
2
3
3
1
2
3
x
It turns out that the double-reflection we just described for g(x) = x 3 is equivalent to rotation around the origin by 180◦ . You can try this equivalence out for yourself by physically
double-reflecting and rotating the book while looking at the preceding graph of g(x) = x 3 .
You can also see the equivalence by using a piece of paper with a smiley-face drawn on the
front: Flipping the paper vertically and then horizontally is equivalent to rotating the paper
by 180 degrees. A function that is preserved under the transformation of 180◦ rotation is
said to have 180◦ rotational symmetry.
These types of symmetries are also called even symmetry and odd symmetry, since
power functions with even powers all have y-axis symmetry and power functions with odd
powers all have rotational symmetry. Because graphical reflections correspond to multiplication by −1 , we can describe functions with even and odd symmetry algebraically as
follows:
DEFINITION 0.9
Even and Odd Functions
A function f is an even function if f (−x) = f (x) for all x in the domain of f .
A function f is an odd function if f (−x) = −f (x) for all x in the domain of f .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
23
Operations, Transformations, and Inverses
For example, the function f (x) = x 2 is even because for all x we have
f (−x) = (−x)2 = x 2 = f (x).
In contrast, the function g(x) = x 3 is odd because for all x we have
g(−x) = (−x)3 = −(x 3 ) = −g(x).
Note that some functions are neither even nor odd; for example h(x) = x 2 + x is one such
function, because h(−x) = (−x)2 + (−x) = x 2 − x, which is equal neither to h(x) nor to
−h(x). Consequently, the function h(x) = x 2 + x has neither y-axis symmetry nor rotational
symmetry.
Inverse Functions
The inverse of a function f is a function that undoes the action of f . For example, the
function that adds 1 to each real number can be undone by subtracting 1 from each real
number. If two functions undo each other, then composing them results in the identity
function. This property suggests the following definition:
DEFINITION 0.10
The Inverse of a Function
If f and g are functions such that
g( f (x)) = x, for all x in the domain of f
f ( g(x)) = x, for all x in the domain of g
then g is the inverse of f and we denote g by f −1 .
Note that the two conditions in this definition guarantee that if a function g is the inverse
of a function f , then f is the inverse of the function g.
√
For example, the functions f (x) = x 3 and g(x) = 3 x are inverses of each other. Intuitively, taking the cube of a number is undone by taking the cube root, and vice versa. It is
easy to verify that these two functions satisfy the condition in Definition 0.10:
3
g( f (x)) = g(x 3 ) = x 3 = x,
√
√
f ( g(x)) = f ( 3 x ) = ( 3 x )3 = x.
CAUTION
1
x
It is important to note that although we use the notation x −1 to denote the reciprocal , the
1
f
notation f −1 does not stand for the reciprocal of f . The notation f −1 used in Definition 0.10
is pronounced “f inverse.” We are now using the same notation for two very different
things, but it should be clear from the context which one we mean.
Not all functions have inverses. For example, consider the squaring function f (x) = x 2
on the domain R. Since f (2) = 4 and f (−2) = 4, an inverse of f would have to send the
input 4 to both 2 and −2; this is clearly not a function. As the following theorem asserts, a
function has an inverse only if no two inputs are ever sent to the same output:
THEOREM 0.11
Invertible Functions are One-to-One
A function f has an inverse if and only if f is one-to-one.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
24
Chapter 0
November 26, 2012
Functions and Precalculus
Since a function is one-to-one if and only if it passes the horizontal line test, this theorem
implies that a function has an inverse if and only if it passes the horizontal line test. The
proof follows directly from the properties of inverse functions:
Proof. We first prove that if a function is invertible, then it must be one-to-one. If f is an invertible
function, then it has an inverse f −1 . Suppose two domain values a and b are sent by f to the same
output f (a) = f (b). Applying f −1 to both sides, we have
f (a) = f (b) =⇒ f −1 ( f (a)) = f −1 ( f (b)) =⇒ a = b.
For the converse, suppose f is one-to-one. Then each element b in the range of f is the output of
exactly one element a from the domain. We can define f −1 (b) to be this element a. Since f is one-toone, this new relationship f −1 will be a function and we will have f −1 (b) = a if and only if f (a) = b.
Therefore if f is one-to-one, then f has an inverse.
The properties of inverses given in the next theorem follow directly from our definition
of an inverse function, that is, from the fact that f −1 undoes the function f . Functions that
have inverses are said to be invertible.
THEOREM 0.12
Properties of Inverses
If f is an invertible function with inverse f −1 , then the following statements hold.
(a) Domain( f −1 ) = Range( f ) and Range( f −1 ) = Domain( f ).
(b) f −1 (b) = a if and only if f (a) = b.
(c) The graph of y = f −1 (x) is the graph of y = f (x) reflected across the line y = x.
Proof. The proofs of parts (a) and (c) are left to Exercises 91 and 92. To prove part (b), suppose
f −1 (b) = a. Applying f to both sides, we have f ( f −1 (b)) = f (a). Since f and f −1 are inverses, their
composition is the identity function. Therefore we have b = f (a), as desired. With an entirely similar
argument we can show that if f (a) = b, then f −1 (b) = a; see Exercise 93.
√
For example, consider the one-to-one function f (x) = x + 1. To find a function
√
that undoes f (x) we solve y = x + 1 for y, obtaining x = y2 − 1. Changing notation so
that x is again the independent variable, we see that the inverse of f (x) is f −1 (x) = x 2 − 1.
√
The following three figures illustrate the properties from Theorem 0.12 for f (x) = x + 1.
√
f (x) = x + 1 with
domain [−1, ∞), range [0, ∞)
y
y
4
4
3
3
4
(2, 3)
3
2
(3, 2)
(0, 1)
1
1
1
1
y
yx
2
2
1
f −1 (x) = x 2 − 1 with
domain [0, ∞), range [−1, ∞)
Reflect f (x) across the line
y = x to obtain f −1 (x)
2
3
4
x
1
(1, 0)
(1, 0)
1
1
1
2
(0, 1)
3
4
x
1
1
2
3
4
x
1
Sometimes a function is not invertible on its largest domain, but is invertible on some
smaller, “restricted” domain. For example, the function g(x) = x 2 − 1 is not invertible on
its usual domain R, but it is invertible on the restricted domain [0, ∞). In general, to find
a restricted domain on which a function is one-to-one, we choose a smaller domain on
which the graph passes the horizontal line test; see Example 4.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
Operations, Transformations, and Inverses
25
Examples and Explorations
EXAMPLE 1
Combinations of functions and their domains
Describe the domains of each of the combinations of f (x) =
follow. Then find an expression for the combination function.
(b) f + g
(a) 3f
(c)
f
g
1
x
and g(x) =
(d) f ◦ g
√
x + 1 that
(e) g ◦ f
SOLUTION
(a) The domain
of 3f is the same as the domain of f , which is x = 0. We have (3f )(x) =
3f (x) = 3
1
x
3
x
= .
(b) The domain of f is x = 0, and the domain of g is x ≥ −1. The domain of their sum f + g
is the intersection of these domains, or [−1, 0) ∪ (0, ∞). For values of x in this domain
√
1
we have ( f + g)(x) = f (x) + g(x) = + x + 1.
x
(c) We have g(x) = 0 only when x = −1, so the quotient
f
g
is defined on the intersection
of the domains of f and g with the point x = −1 removed, or (−1, 0) ∪ (0, ∞). On this
domain we have
1
f (x)
1/x
f
= √
.
(x) =
= √
g
g(x)
x+1
x x+1
(d) For a value x to be in the domain
√ of f ◦ g, it must first be in the domain [−1, ∞) of
g.
√ Then the value of g(x) = x + 1 must be in the domain of f , so we must have
x + 1 = 0, or in other words, x = −1. Therefore the domain of f ◦ g is (−1, ∞). For
√
1
values of x in this domain we have ( f ◦ g)(x) = f ( g(x)) = f ( x + 1 ) = √
. Notice
that this equation is consistent with our calculation of the domain.
x+1
(e) For a value x to be in the domain of g◦f , it must first be in the domain (−∞, 0)∪(0, ∞) of
1
1
f . Then the value of f (x) = must be in the domain of g, so we must have ∈ [−1, ∞).
x
1
x
x
≥ −1 when x ≤ −1 or x > 0, the domain of g ◦ f is (−∞, −1] ∪ (0, ∞). For
1
1
=
+ 1. Again notice
values of x in this domain we have ( g ◦f )(x) = g( f (x)) = g
Since
x
x
that the domain that we found does make sense with this equation.
EXAMPLE 2
Vertical and horizontal translations, stretches, and reflections
The figure that follows shows a piece of the graph of f (x) = 3 + 2x − x 2 with five marked
points. Find equations and graphs for each of the given transformations. On each graph,
mark the new coordinates of the five marked points.
y
8
6
4
2
6 4 2
2
4
2
4
6
x
(a) f (x) + 3, f (x) − 3, f (x + 3), and f (x − 3)
1
1
(b) 2 f (x), f (x), f (2x), and f x
2
2
(c) −f (x) and f (−x)
6
8
10
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
26
Chapter 0
November 26, 2012
Functions and Precalculus
SOLUTION
(a) The equations for the four transformations are
f (x) + 3 = (3 + 2x − x 2 ) + 3 = 6 + 2x − x 2 ,
f (x) − 3 = (3 + 2x − x 2 ) − 3 = 2x − x 2 ,
f (x + 3) = 3 + 2(x + 3) − (x + 3)2 = −4x − x 2 , and
f (x − 3) = 3 + 2(x − 3) − (x − 3)2 = −12 + 8x − x 2 .
The graphs of the transformations are shifts up, down, left, and right, respectively, of
the original graph by 3 units, as shown in the red graphs in the following figures:
y = f (x) + 3
y = f (x) − 3
y
6 4 2
y = f (x + 3)
y
y = f (x − 3)
y
y
8
8
8
8
6
6
6
6
4
4
4
4
2
2
2
2
2
4
4
6
x
6 4 2
2
2
4
4
6
x
6 4 2
2
4
2
4
6
x
2
6 4 2
2
2
4
6
6
6
8
8
8
8
10
10
10
10
4
6
x
6
In each case, every point on the original graph is shifted in some direction by 3 units.
For example, the point (1, 4) on the original graph becomes (1, 7) on the graph of
f (x) + 3, (1, 1) on the graph of f (x) − 3, (−2, 4) on the graph of f (x + 3), and (4, 4)
on the graph of f (x − 3). The other four marked points move in a similar fashion.
(b) Algebraically, the four transformations are given by the equations
2 f (x) = 2(3 + 2x − x 2 ) = 6 + 4x − 2x 2 ,
1
1
3
1
f (x) = (3 + 2x − x 2 ) = + x − x 2 ,
2
2
2
2
f (2x) = 3 + 2(2x) − (2x)2 = 3 + 4x − 4x 2 , and
2
1
1
1
1
x = 3 + 2 x − x = 3 + x − x 2.
f
2
2
2
4
The first two transformations cause the graph of f to stretch or compress vertically, and
the last two cause the graph to compress or stretch horizontally, as shown in the red
graphs in the next four figures.
y = 2 f (x)
y=
y
6 4 2
1
2
y = f (2x)
f (x)
y
y=f
y
8
8
8
6
6
6
6
4
4
4
4
2
2
2
4
2
4
6
6 4 2
2
4
2
4
6
x
2
6 4 2
2
4
1
x
2
y
8
x
2
4
6
x
2
4 2
2
4
6
6
6
8
8
8
8
10
10
10
10
CONFIRMING PAGES
6
2
4
6
8
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
Operations, Transformations, and Inverses
27
In the first two transformations it is the y-coordinate that changes; for example, the
point (1, 4) on the original graph moves to (1, 8) for the first transformation and to
(1, 2) for the second. For the last two
it is the x-coordinate that
transformations,
1
changes; for example, (1, 4) moves to , 4 for the third transformation and to (2, 4)
2
for the fourth. The other marked points move in a similar fashion, with either the
x- or the y-coordinate being multiplied or divided by 2.
(c) The equations for the two transformations are
−f (x) = −(3 + 2x − x 2 ) = −3 − 2x + x 2 and
f (−x) = 3 + 2(−x) − (−x)2 = 3 − 2x − x 2 .
The first transformation gives a vertical reflection across the x-axis, with each marked
point (x, y) moving to the point (x, −y), as shown next at the left. The second transformation causes a horizontal reflection across the y-axis, with each marked point (x, y)
moving to the point (−x, y), as shown at the right.
y = −f (x)
y = f (−x)
y
y
8
8
6
6
4
4
2
6 4 2
2
4
2
4
6
2
x
6 4 2
6
EXAMPLE 3
2
4
2
4
6
x
6
8
8
10
10
Testing if functions are even or odd
Determine whether each of the following functions is even, odd, or neither:
(a) f (x) =
1
x
(b) g(x) = x 4 − x 2
(c) h(x) =
2+x
1 + x2
SOLUTION
(a) To determine whether f is even or odd (or neither) we must calculate f (−x) and determine if it is equal to f (x), −f (x), or neither. We have
f (−x) =
so f (x) =
1
x
1
1
= − = −f (x),
−x
x
is an odd function.
(b) The function g(x) is even because
g(−x) = (−x)4 − (−x)2 = x 4 − x 2 = g(x).
(c) Since we have
h(−x) =
2−x
2 + (−x)
=
,
2
1 + (−x)
1 + x2
which is equal neither to h(x) nor to −h(x), this function is neither even nor odd.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
28
Chapter 0
CHECKING
THE ANSWER
November 26, 2012
Functions and Precalculus
The graphs of the functions f , g, and h are shown next. Note that f has rotational symmetry
about the origin, g has y-axis reflectional symmetry, and h does not have either type of
symmetry.
f (x) =
1
x
g(x) = x 4 − x 2
is an even function
is an odd function
y
3
3
2
2
2
1
3 2 1
1
1
2
3
x
3
1
1
2
2
2+x
1 + x2
is neither even nor odd
y
y
3
EXAMPLE 4
h(x) =
1
1
2
x
3 2 1
1
1
2
3
x
2
3
1
Graphically finding a restricted domain on which a function is invertible
Explain why the function f graphed here is not invertible on its domain. Then find three
restricted domains on which the function does have an inverse.
y
4
3
2
1
3 2 1
1
1
2
3
x
2
SOLUTION
This function is not invertible on the domain R, because it does not pass the horizontal
line test and therefore fails to be a one-to-one function. However, small enough pieces
of the graph of f do pass the horizontal line test and therefore have an inverse on that
restricted domain. The three graphs that follow show the graph of f on the restricteddomain domains [−1, 1], (−∞, −1], and [1, ∞), respectively. All three of these restricted
graphs pass the horizontal line test and thus are invertible.
y = f (x) on [−1, 1]
y = f (x) on (−∞, −1]
y
y = f (x) on [1, ∞)
y
y
4
4
4
3
3
3
2
2
2
1
3 2 1
1
2
1
1
2
3
x
3 2 1
1
1
1
2
3
x
3 2 1
1
2
2
CONFIRMING PAGES
1
2
3
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
EXAMPLE 5
29
Operations, Transformations, and Inverses
The graphical and algebraic relationships of inverse functions
Consider the function f (x) = 2x − 2.
(a) Explain why f must have an inverse. Then sketch graphs of y = f (x) and y = f −1 (x)
and explain the relationship between these two graphs.
(b) Find a formula for f −1 (x).
(c) Algebraically verify that f ( f −1 (x)) = x and f −1 ( f (x)) = x.
SOLUTION
(a) f (x) = 2x − 2 has an inverse because it is a non-constant linear function and thus
passes the horizontal line test and is one-to-one. By reflecting the graph of
y = f (x) across the line y = x, we obtain the graph of this inverse; see the following figures:
y = f −1 (x)
reflect across line y = x
y = f (x)
y
y
y
4
4
4
3
3
3
2
2
2
1
1
1
2 1
1
1
2
3
4
x
2 1
1
2
1
2
3
x
4
2 1
1
1
2
3
4
x
2
2
Whenever (x, y) is a point on the graph of f , the point (y, x) is on the graph of f −1 . For
example, in the preceding graphs we see that (2, 2) and (3, 4) are on the graph of f
while (2, 2) and (4, 3) are on the graph of f −1 .
(b) We could find a formula for f −1 by using the two-point form of a line and the two
points shown in the rightmost figure. However, we will find it by solving the equation
y = 2x − 2 for x:
1
2
1
2
y = 2x − 2 =⇒ y + 2 = 2x =⇒ x = ( y + 2) = y + 1.
If y = f (x), then x = f −1 ( y); thus we have shown that f −1 ( y) =
f −1
1
y
2
+ 1. Since we
would rather represent the independent variable of
by the traditional letter x,
1
replace the y’s in the equation with x’s to get f −1 (x) = x + 1. Notice that this
2
equation does appear reasonable, given the slope and y-intercept in the rightmost
graph.
1
2
(c) For f (x) = 2x − 2 and f −1 (x) = x + 1, we have
f ( f −1 (x)) = f
1
1
x + 1 = 2 x + 1 − 2 = x + 2 − 2 = x,
2
2
1
2
f −1 ( f (x)) = f −1 (2x − 2) = (2x − 2) + 1 = x − 1 + 1 = x.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
30
Chapter 0
November 26, 2012
Functions and Precalculus
When we write ( f + g)(x) = f (x) + g(x), what kind of objects is the first “+” symbol
TEST YOUR
? UNDERSTANDING
adding together? What kind of objects is the second “+” adding together?
Why does it make sense that adding or multiplying a constant to the dependent variable
of y = f (x) would cause a vertical change in the graph?
Why does it make sense that the graph of y = f (x − 5) would be obtained by shifting
the graph of y = f (x) to the right, rather than to the left?
In Definition 0.10 we talk about the inverse of f . If a function has an inverse it has only
one. Can you explain why?
What is the difference between the inverse and the reciprocal of the function f (x) =
2x − 2 from Example 5? Why might someone wrongly confuse these two functions?
EXERCISES 0.2
Thinking Back
Evaulating functions: Find each of the following evaluations of
x
:
the function f (x) =
x−1
f (1.5)
f
1
x
Solving equations: Each equation that follows expresses s in
terms of r. Solve for r (i.e., write r in terms of s).
f (x 2 )
f (x 2 + 1)
s=
r−2
3
s=
3
r−2
f ( f (x))
f ( f (x 2 ))
s=
r+1
r
s=
r+2
r+3
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The domain of a function of the form
f + g is equal to the domain of f or the domain of g,
whichever is smaller.
(b) True or False: The domain of a function of the form f ·g
is equal to the intersection of the domains of f and g.
(c) True or False: If the graph of y = f (x) contains the
point (a, b), then the graph of y = f (x) + C must contain the point (a, b + C).
(d) True or False: If the graph of y = f (x) contains the
point (a, b), then the graph of y = f (x + C) must contain the point (a + C, b).
(e) True or False: The inverse of the one-to-one function
f (x) = x 5 is f −1 (x) = x −5 .
(f) True or False: If f is an invertible function, then
1
f −1 = .
f
(g) True or False: Every even function is a function that
involves only even exponents.
(h) True or False: If f is an even function and (0, b) is a
point on the graph of y = f (x), then (0, −b) must also
be on the graph of y = f (x).
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A pair of functions f and g for which f ◦ g happens to
be equal to g ◦ f .
(b) A function f that is its own inverse.
(c) A function f that is both even and odd.
3. Explain what the definition of ( f − g)(x) ought to be.
Show that this definition is just a combination of the
definitions of ( f + g)(x) and k f (x).
4. Suppose (2, 5) is a point on the graph of y = f (x). Fill in
the blanks with the transformed coordinates of this point
under each of the following transformations:
(a)
is on the graph of f (x) − 4.
(b)
is on the graph of f (x − 4).
(c)
(d)
is on the graph of −7f (x).
1
is on the graph of f x .
(e)
is on the graph of 3f (x + 1).
(f)
is on the graph of f (3x + 1).
3
5. Fill in the blanks as appropriate. There may be more than
one possible answer.
is on the graph of y = f (x),
(a) If the point
then (4, 2) is on the graph of y = f (x − 3).
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
(b) If (3, −2) is on the graph of y = f (x), then (6, −2) is
.
on the graph of the function
(c) If (1, 4) is on the graph of an even function y = f (x),
is also on the graph of y = f (x).
then
(d) If (−2, 5) is on the graph of an odd function y = f (x),
is also on the graph of y = f (x).
then
6. Suppose f has domain [1, ∞) and g has domain [−4, 4].
Suppose also that f (x) is nonzero except at x = 2 and x = 5
and that g(x) is nonzero except at x = −1 and x = 1. Find
the domains of the following functions (if possible):
(a) 3 f + 4 g
(b)
1
fg
(c)
9. Given that y = f (x) has the graph on the left, use transformations to find a formula in terms of f (x) for the function
graphed on the right.
A transformation of y = f (x)
y
y
2
2
1
y
(a)
4
3
3
2
2
1
1
3 2 1
1
1
2
x
3
3 2 1
1
2
3
1
2
3
x
2
2
y
(c)
1
y
(d)
3
2
2
1
2 1
1
1
x
2
3 2 1
1
1
x
2
2
3
x
-3
f (x)
4
-2
-1
0
1
2
-2
1
3
14. Suppose f is a function with domain R whose right-hand
side is as shown here. Sketch the left-hand side of the
graph so that (a) f is an even function, (b) f is an odd
function, and (c) f is neither even nor odd.
1
3 2 1
1
1
2
3
x
3 2 1
1
1
2
3
x
y
4
2
2
3
3
3
2
4
4
1
10. Given that y = f (x) has the graph on the left, use transformations to find a formula in terms of f (x) for the function
graphed on the right.
y = f (x)
A transformation of y = f (x)
y
y
2
1
1
2
3
x
3 2 1
1
2
2
3
3
4
4
1
2
3
3 2 1
1
1
2
3
x
2
15. Suppose f is an invertible function with inverse f −1 . What
is ( f −1 )−1 ? Explain your answer.
16. Given that f is an invertible function, fill in the blanks.
2
1
3 2 1
1
y
(b)
4
13. Complete the entries in the following table two ways:
(a) to make an even function and (b) to make an odd
function:
1
, what is f (x)?
x2 + 1
1
(b) If h(x) = x 2 − 1 and h(l(x)) = 4 − 1, what is l(x)?
x
1
1
(c) If u(x) =
and y(u(x)) =
, what is y(x)?
1−x
1−x
(a) If g(x) = x 2 and f ( g(x)) =
y = f (x)
11. If f (0) = 2, can f be an odd function? What if f (0) is
undefined? Explain your answers.
12. Determine graphically whether each of the following four
functions is even, odd, or neither.
1
f +g
7. Suppose f is a function with domain [2, ∞) and range
[−3, 3], and let g be a function with domain [−10, ∞) and
range [0, ∞).
(a) What is the domain of the composition g ◦ f ? Justify
your answer.
(b) It is not possible to determine the domain of f ◦ g in
this example; explain why not. What extra information would you have to know to be able to determine
the domain of f ◦ g?
(c) Is there enough information here to determine the
domain of the composition f ◦ f ? What about the
function g ◦ g?
8. Use compositions to answer each of the following:
31
Operations, Transformations, and Inverses
x
(a) If f (−1) = 0, then f −1 (0) =
(b) If (2, 3) is on the graph of f , then
graph of f −1 .
.
is on the
(c) If
is on the graph of f , then (−2, 4) is on
the graph of f −1 .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
32
Chapter 0
November 26, 2012
Functions and Precalculus
17. If an invertible function f has domain [−1, 1) and range
(−∞, 3], what are the domain and range of f −1 ? Sketch
the graph of a function f with the domain and range
given, and on the same set of axes sketch the graph of
its inverse f −1 .
18. Consider the function f (x) = x 5 + 4x 3 + 2x + 1. Alina
says that f is invertible, because its graph appears to be
one-to-one. Linda says that f is not invertible, because
she cannot figure out how to solve y = x 5 + 4x 3 + 2x + 1
for x in terms of y. Who is correct, and why?
20. Use the values given in the table to fill in the missing
values. There is only one correct way to fill in the table.
0
f (x)
1
g(x)
0
1
1
2
3
4
4
3
1
2
g(x)
3
1
2
h(x)
2
( f ◦ g)(x)
4
(h ◦ f )(x)
19. Use the values given in the table to fill in the missing
values. There is only one correct way to fill in the table.
x
x
f (x)
4
3
2
( g ◦ g)(x)
( g ◦ f ◦ h)(x)
2
2
−2
h(x)
1
(−2f )(x)
4
( f + 2g − h)(x)
4
( fg)(x)
g
(x)
2
−1
3
h
Skills
√
1
, and h(x) = x, find
x−2
Given that f (x) = x 2 + 1, g(x) =
the domain of each function in Exercises 21–29. Then find an
equation for the function and calculate its value at x = 1.
g
(x)
21. ( f + g)(x)
22. (3f h)(x)
23.
h
24. ( g ◦ f )(x)
25. ( g ◦ g)(x)
26. ( g ◦ h ◦ f )(x)
27. g(x − 5)
28. h(3x) + 1
29. h(3x + 1)
The table that follows defines three functions f , g, and h.
Create additional rows for the table for each function in
Exercises 30–38. (For some transformations, you may have to
use different x-values than the ones in the table.)
x
0
1
2
3
4
5
6
f (x)
0
1
3
2
3
0
2
g(x)
1
0
1
1
0
1
0
h(x)
3
2
0
3
2
3
1
30. ( f − g)(x)
31. 2f (x) + 3
32. ( gh)(x)
33. (h ◦ g)(x)
34. ( g ◦ h)(x)
35. ( f ◦ f ◦ f )(x)
36. −h(−x)
37. g(x − 1)
38. f (2x)
Use the graphs of f and g given here to sketch the graphs
of the functions in Exercises 39–50. Label at least four points
on each graph. Don’t find or use equations for the given
graphs.
y = f (x)
y = g(x)
y
y
3
3
2
2
1
1
3 2 1
1
1
2
3
x
3 2 1
1
2
2
3
3
1
2
3
1 x
39. ( f + g)(x)
40. 2 − 3f (x)
41. g
42. g(x − 2) + 1
43. (−.5f )(x)
44. ( fg)(x)
45. ( f ◦ g)(x)
46. ( g ◦ f )(x)
47. ( f ◦ f )(x)
48. g −1 (x)
49.
1
f
(x)
CONFIRMING PAGES
50.
2
1
g
(x)
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.2
Suppose f and g are the piecewise-defined functions defined
here. For each combination of functions in Exercises 51–56,
(a) find its values at x = −1, x = 0, x = 1, x = 2, and x = 3,
(b) sketch its graph, and (c) write the combination as a
piecewise-defined function.
2x + 1, if x ≤ 0
−x, if x < 2
g(x)
=
f (x) =
5, if x ≥ 2
x 2 , if x > 0
51. ( f + g)(x)
52. 3 f (x)
53. ( g ◦ f )(x)
54. f (x) − 1
55. f (x − 1)
56. g(3x)
Find two nontrivial ways to write each of the functions in
Exercises 57–60 as a composition f = g ◦ h. “Nontrivial”
means that you should not choose g(x) = x or h(x) = x.
57. f (x) = 3x + 5
58. f (x) = (3x + 5)
2
59. f (x) =
6
x+1
2
1+
2
64. f (x) = − 3
5x
1
66. f (x) =
x+1
65. f (x) =
3
x
x2 + 1
y
71.
y
72.
5
5
4
4
3
3
2
2
1
1
3 2 1
1
1
2
3
x
y
3 2 1
1
2
3
4
3
2
1
3
2
1
3 2 1
1
1
2
3
x
2
2
Use Definition 0.10 to show that each pair of functions in
Exercises 67–70 are inverses of each other.
2
1
67. f (x) = 2 − 3x and g(x) = − x +
3
3
√
68. f (x) = x 2 restricted to [0, ∞) and g(x) = x
x
x
and g(x) =
69. f (x) =
1−x
1+x
1
1
and g(x) =
70. f (x) =
2x
2x
1
x
y
74.
4
x2
Determine algebraically whether the functions in Exercises 61–66 are even, odd, or neither. Afterwards, verify your
answers by inspecting the symmetries of the graphs.
62. f (x) = 1 − 2x
61. f (x) = x 4 + 1
63. f (x) = x 3 + x 2
For each invertible function in Exercises 71–74, sketch the
graph of f −1 and label three points on its graph. If a function
is not invertible, then find a restricted domain on which it is
invertible and sketch a graph of the restricted inverse. Don’t
find or use equations for the given graphs.
73.
x2
60. f (x) = √
33
Operations, Transformations, and Inverses
1 1
2
3
4
1
2
x
Find the inverse of each function in Exercises 75–80. Then
check your answers by graphing f and f −1 .
75. f (x) =
1 − 5x
2
1
x
78. f (x) =
2
x+1
x−1
x+1
80. f (x) =
x−2
x
77. f (x) = 1 +
79. f (x) =
76. f (x) = 1.2 − 3.5x
Applications
81. CarpetKing charges $4.25 per square foot for its Deluxe
ThriftySoft carpet, plus a flat fee of $200.00 for delivery
and installation.
(a) If a square room measures x feet on a side, and S is
the number of square feet of floor in the room, write
down S as a function of x.
(b) Write down a function that describes the cost C of
carpeting a room enclosing S square feet.
(c) Write down a function that describes the cost C of
carpeting a square room that measures x feet on a
side. Explain how this function is a composition.
82. The first table that follows shows the number of deer in
Happyland Forest Park during 1990–1995. The number of
deer seems to affect the number of ticks in the park; the
second table shows the number of ticks that can be expected for various numbers of deer in the park.
Year 1990 1991 1992 1993 1994 1995
Deer
183
180
177
179
184
181
Deer 177 178 179 180 181 182 183 184
Ticks 850 855 860 865 870 875 880 885
(a) Use the tables to estimate the number of ticks that
were in the park in 1995.
(b) Make a table that predicts the number of ticks in the
park during 1990–1995.
(c) Explain how your table represents a composition of
the deer and tick tables.
83. Juri’s custom printing t-shirt shop charges $12.00 per
t-shirt plus a $20.00 setup fee.
(a) Write a function that describes the cost C of printing
n t-shirts at Juri’s store.
(b) Show that C(n) is an invertible function.
(c) Find the inverse of C(n). What does this inverse function n(C) represent in terms of C and n?
(d) Use the function you found in part (c) to calculate the
number of shirts you can produce with $150.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
34
Chapter 0
November 26, 2012
Functions and Precalculus
84. The math teachers at Pinnacle High School have discovered a relationship between the number of hours their
students watch television each day and their students’
scores on math tests. The following table shows this
relationship:
Hours of T.V.
0
1
2
3
4
5
6
Grade on test 92 86 80 74 68 62 56
(b) Show that your function g(t) is invertible.
(c) Find the inverse of g(t). What does this function t( g)
represent in terms of t and g?
(d) Use t( g) to predict the number of hours that Calvin
watched television each day given that he scored a 40
on his math test. How much television can he be
allowed to watch each day if his mother wants him
to get an 85 on his next test?
(a) Find a linear function that describes the grade g on
a student’s math test as a function of the number
of hours t of television that the student watches
every day.
Proofs
85. Prove algebraically that if f (x) = x k , where k is a positive
integer, then the graphs of y = f (2x) and y = 2k f (x) are
the same.
86. Prove that if (x, y) is a point on the graph of y = f (x) and
C is a real number, then
(a) (x, y + C) is a point on the graph of y = f (x) + C.
(b) (x − C, y) is a point on the graph of y = f (x + C).
87. Prove that if (x, y) is a point on the graph of y = f (x) and
k = 0, then
(a) (x, ky) is a point on the graph of y = k f (x).
1
(b)
x, y is a point on the graph of y = f (k x).
k
88. Prove that if (x, y) is a point on the graph of y = f (x),
then
(a) (x, −y) is a point on the graph of y = −f (x).
(b) (−x, y) is a point on the graph of y = f (−x).
89. Prove that every odd function that is defined at x = 0
must pass through the origin.
90. Prove that the algebraic definitions of even and odd
functions imply even and odd graphical symmetry,
by showing that:
(a) If (x, y) is a point on the graph of an even function
y = f (x), then (−x, y) is also on the graph.
(b) If (x, y) is a point on the graph of an odd function
y = f (x), then (−x, −y) is also on the graph.
91. Prove Theorem 0.12 (a) by using Definition 0.10 to argue
that the domain of f −1 is the range of f and that the range
of f −1 is the domain of f .
92. Prove Theorem 0.12 (c) by showing that if f is a function
with an inverse f −1 and (x, y) is on the graph of y = f (x),
then (y, x) is on the graph of y = f −1 (x). Why does this
conclusion imply that the graph of f −1 (x) is the reflection
of the graph of f (x) across the line y = x?
93. Suppose f is invertible with inverse f −1 . Prove that if
f (a) = b, then f −1 (b) = a.
94. Prove that an invertible function f can have only one inverse. (Hint: Suppose g and h are both inverses of a function f ,
and suppose also that there is some real number x0 for which
g(x0 ) = h(x0 ). Show that these suppositions together produce
a contradiction.)
Thinking Forward
Transformations of trigonometric graphs: Consider the function
f (x) = sin x shown in the following figure:
f (x) = sin x
y
π
2
π
3π
2
2π
x
1
f (x)
especially careful when considering what happens at
the places where f (x) has zeros. (The resulting graph
will be the graph of csc x, so this problem illustrates
1
2π 3π π π
2
2
Use transformations to sketch the graph of the function f (x + 2π ). What do you notice about this new
graph? What does this mean about the function f (x) =
sin x?
1
Sketch the graph of the reciprocal
of f (x). Be
that
1
= csc x.)
sinx
The function f (x) = sin x is not one-to-one and therefore is not invertible. However,
f (x) =
sin x is invertπ π
. Sketch this re2 2
ible on the restricted domain − ,
Use transformations
to sketch the graph of the func
π
. (The resulting graph will be the
tion f x +
2
graph
of
cos
x,
so
this problem illustrates the identity
sin x +
π
2
stricted function, and then sketch its inverse f −1 (x).
(The graph of f −1 (x) is the graph of the function known
as sin−1 x.)
= cos x.)
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.3
0.3
35
Algebraic Functions
ALGEBRAIC FUNCTIONS
Power functions, polynomial functions, and rational functions
Functions that involve absolute values
Domains, properties, and graphs of algebraic functions
Power Functions
A function is algebraic if it can be expressed in terms of constants and a variable x by using
only arithmetic operations (+, −, ×, and ÷) and rational constant powers of the variable.
For example,
√
√
3x 2 − 5
f (x) =
and g(x) = ( x + 3 x )7/2
1−x
are algebraic functions, but h(x) = 2 x and k(x) = sin x are not.All algebraic functions are
combinations and/or compositions of power functions:
DEFINITION 0.13
Power Functions
A power function is a function that can be written in the form
f (x) = Ax k
for some nonzero real number A and some rational number k.
In this definition, the constant k is called the power or exponent, and the constant A is
called the coefficient. Note that the exponent k must be a rational number and the variable must be in the base; this means that, for example, f (x) = x π and f (x) = 10 x are not
considered power functions.
Although power functions all have the simple form f (x) = Ax k , they vary greatly
for different values of exponent k. The following graphs illustrate eight common power
functions:
y = x2
y
y
4
2
3
1
2
2
1
2
1
2
x
y = x 1/2
4
2
3
1
1
1
2
x
3 2 1
1
1
2
3
x
1
3
3 2 1
y = x 2/3
y = x 1/3
y
y
3
4
x
2
x
y
2
1
2
3
x
1
1
1
2
3
2
1
1
2
3
2
3 2 1
1
y = x 3/2
2
1
2
2
2
y
y
y
3
1
1
y = x −2
y = x −1
y = x3
2
1
1
2
x
CONFIRMING PAGES
1
2
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
36
Chapter 0
November 26, 2012
Functions and Precalculus
Although all eight of these graphs represent power functions, they are very different. They
have different domains and ranges, some have asymptotes and some do not, some have
sharp corners and some do not, and so on. However, all power functions with integer
powers or positive rational powers have graphs that look similar to one of these eight
graphs. For example, if k is a positive odd integer, then the graph of f (x) = x k looks a lot
p
p
like the graph of x 3 . If is a positive rational number with p even and q odd, and if < 1,
q
q
then the graph of g(x) = x p/q looks a lot like the graph of x 2/3 .
Polynomial Functions
A polynomial function is a finite sum of power functions that have nonnegative integer
powers. The following definition provides a general notation:
DEFINITION 0.14
Polynomial Functions
A polynomial function of degree n is a function that can be written in the form
f (x) = a n x n + a n−1 x n−1 + a n−2 x n−2 + · · · + a2 x 2 + a1 x + a 0
for some integer n ≥ 0 and some real numbers a 0 , a1 , . . . a n with a n = 0.
As a matter of convention, we also say that the constant zero function f (x) = 0 is a
polynomial function (and that its degree is undefined).
The numbers a i (for i = 0, 1, 2, . . . , n) are called the coefficients of the polynomial. Note
that the coefficient belonging to the term containing the power x i is conveniently named
a i ; for example, the coefficient of the x 2 term is called a 2 . The coefficient a n belonging to
the highest power of x is called the leading coefficient, and the term a n x n containing the
highest power of x with a n = 0 is called the leading term. The coefficient a 0 is called the
constant term.
We have special names for some polynomials according to their degrees. For example,
polynomials of degrees 0, 1, 2, 3, 4, and 5 are called constant, linear, quadratic, cubic,
quartic, and quintic polynomials, respectively. (Note: We will require non-zero leading
coefficients for quadratics and higher degrees, but not for linear functions; in other words,
we will consider constant functions to be linear.) Higher degrees can sometimes result in
more roots and more turning points in the graph of a polynomial; for example, examine
the following cubic, quartic, and quintic polynomials:
f (x) = x 3 + x 2 − 2x
= x(x − 1)(x + 2)
g(x) = x 4 + x 3 − 2x 2
= x 2 (x − 1)(x + 2)
h(x) = x 5 − x 4 − 4x 3 + 4x 2
= x 2 (x − 1)(x + 2)(x − 2)
y
y
y
10
10
10
5
5
5
3 2 1
5
1
2
3
x
3 2 1
1
2
3
x
3 2 1
5
5
CONFIRMING PAGES
1
2
3
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.3
Algebraic Functions
37
A quadratic expression is irreducible if it cannot be factored with real-number coefficients; that is, ax 2 + bx + c is irreducible if it cannot be written in the form a(x − r 1 )(x − r 2 )
for some real numbers r 1 and r 2 . A quadratic expression ax 2 + bx + c is irreducible if and
only if its discriminant b 2 − 4ac is negative; think about the quadratic formula to see why.
For example, the quadratic expressions x 2 + 5 and x 2 + x + 7 are irreducible.
Note that every linear factor (x − r) of a polynomial corresponds to a root x = r of
that polynomial, since if a polynomial f (x) can be factored as f (x) = (x − r)g(x) for some
other polynomial g(x), then f (r) = (r − r)g(r) = 0 and thus x = r is a root of f (x). It happens
that every polynomial function can be factored into linear factors (which correspond to
real-number roots) and/or irreducible quadratic factors (which do not correspond to realnumber roots). However, just because a polynomial has a factorization doesn’t mean that
we have an easy way to actually factor that polynomial!
The next theorem describes four key graphical properties of polynomial functions. The
first part of Theorem 0.15 is related to the Fundamental Theorem of Algebra, and the proof
of this deep theorem is beyond the scope of this course. We will have the tools to prove the
last three parts of Theorem 0.15 once we study limits and derivatives.
THEOREM 0.15
Graphical Properties of Polynomial Functions
If f is a polynomial function of degree n, then the graph of f
(a) has at most n real roots;
(b) has at most n − 1 local extrema;
(c) is “smooth” and “unbroken” on R and has no asymptotes;
(d) behaves like the graph of its leading term at the “ends” of the graph.
The last part tells us that a polynomial function f (x) = a n x n + a n−1 x n−1 + · · · + a1 x + a 0
with a n = 0 will behave like the power function a n x n at its “ends.” This means that the
“ends” of the graph of a polynomial always looks like one of the four graphs that follow,
depending on whether the degree n is even or odd and whether the leading coefficient an is
positive or negative. The dashed part of each graph indicates that this part of the theorem
does not tell us the behavior in the middle of the graph, only at the ends.
n even, an > 0
n even, an < 0
n odd, an > 0
y
y
x
n odd, an < 0
y
x
y
x
x
Rational Functions
A rational number is a number that can be written as a quotient of the simplest possible
numbers you can imagine, namely, integers. Similarly, a rational function is a quotient of
very simple functions, namely, polynomials.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
38
Chapter 0
DEFINITION 0.16
November 26, 2012
Functions and Precalculus
Rational Functions are Quotients of Polynomials
A rational function is a function that can be written as the quotient of two polynomial
functions
p(x)
a n x n + an−1 x n−1 + · · · + a1 x + a 0
f (x) =
,
=
q(x)
bm x m + bm−1 x m−1 + · · · + b1 x + b 0
for any x such that q(x) = 0.
Graphs of rational functions are highly dependent on small changes in their numerators
and denominators; for example, examine the three following rational functions:
f (x) =
(x − 3)
(x − 3)(x − 1)
y
1
(x − 3)2
(x − 3)(x − 1)
g(x) =
h(x) =
y
y
4
5
4
3
4
3
2
3
2
1
2
1
1
1
2
3
4
x
1
1
2
1
(x − 3)2
(x − 3)3 (x − 1)
1
2
3
4
x
1
1
1
2
3
4
x
2
3
2
3
4
3
4
We say that the graph of a function f has a hole at x = c if the graph of f is a simple unbroken
curve near x = c, but at x = c there is a point missing from the graph. We will have a more
exact
definition of holes in Chapter 1. For example, the first graph shown has a hole at
3,
1
2
, and the second graph has a hole at (3, 0).
As the preceding examples suggest, the domain, roots, and holes in the graph of a
rational function are determined by the roots of its numerator and denominator. These
relationships are made precise in Theorems 0.17 and 0.18.
THEOREM 0.17
Graphical Properties of Rational Functions
p(x)
If f (x) =
is a rational function, then the following are true.
q(x)
(a) f is not defined at the roots of q(x).
(b) f has roots at the points that are roots of p(x) but not roots of q(x).
(c) f has holes at the points that are roots of both p(x) and q(x), provided that the roots
have higher or equal multiplicity in p(x) than in q(x).
The multiplicity of a root x = c, as mentioned in the theorem, is the number of times that
(x − c) is a factor of the numerator or of the denominator. The first two parts of this theorem follow directly from properties of polynomials, roots, and quotients; see Exercises 92
and 93. The proof of the third part of the theorem will have to wait until we cover limits in
Chapter 1.
The next theorem describes the asymptotes of rational functions in terms of their
numerator and denominator polynomials. Its proof is necessarily postponed until we
have studied limits in Chapter 1.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.3
THEOREM 0.18
Algebraic Functions
39
Vertical and Horizontal Asymptotes of Rational Functions
p(x)
Suppose f (x) =
is a rational function in which p(x) has degree n and q(x) has
q(x)
degree m.
(a) f has vertical asymptotes at the roots of q(x), provided that the roots have a higher
multiplicity in q(x) than in p(x).
(b) If n < m, then f has a horizontal asymptote at y = 0.
(c) If n = m, then f has a horizontal asymptote at y =
an
,
bn
the ratio of the leading
coefficients of p(x) and q(x).
(d) If n > m, then the graph of f does not have any horizontal asymptotes.
For example, looking back at our three graphs of rational functions, we see that the first and
third functions are “bottom heavy” (since the degree of the denominator is greater than
that of the numerator) and have horizontal asymptotes at y = 0. The second is “balanced”
(since the degrees of the numerator and denominator are the same) and has a horizontal
asymptote at y = 1.
Absolute Value Functions
Recall that the absolute value |x| of a real number x is equal to x if x is positive (or zero) and
is equal to −x if x is negative. For example, |2| = 2 while |−2| = −(−2) = 2. Thus we can
write the function f (x) = |x| as a piecewise-defined function by splitting the definition of
|x| into two cases: x ≥ 0 and x < 0.
DEFINITION 0.19
The Absolute Value Function
The absolute value function is the piecewise-defined function |x| =
x, if x ≥ 0
−x, if x < 0.
The graph of y = |x| is a combination of the graph of y = x on [0, ∞) and the graph of
y = −x on (−∞, 0):
y = |x|
4
3
2
1
−4 −3 −2 −1
−1
1
2
3
4
Of course, in general we might wish to take the absolute value of a more complicated
expression. In the more general case we do exactly the same thing: The absolute value will
leave positive quantities untouched, but flip the sign of negative quantities.
DEFINITION 0.20
The Absolute Value of a Function
The absolute value of a function g(x) is |g(x)| =
g(x), for all x with g(x) ≥ 0
−g(x), for all x with g(x) < 0.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
40
Chapter 0
December 10, 2012
Functions and Precalculus
For example, consider the function g(x) = 2x − 1 and its absolute value f (x) = |g(x)| =
|2x − 1|. When 2x − 1 is positive or zero (i.e., when x ≥ 1/2) the absolute value remains
2x−1. But when 2x−1 is negative (i.e., when x < 1/2) the absolute value of g(x) is −(2x−1).
The graph of y = |2x − 1| is a combination of the graphs of y = 2x − 1 and y = −(2x − 1),
switching between graphs at x = 1/2, as shown in the following figures:
y = 2x − 1
y = −(2x − 1)
y = |2x − 1| =
y
y
y
3
3
3
2
2
2
1
1
1
⫺1 ⫺0.5
⫺1
0.5
1
1.5
x
2
⫺1 ⫺0.5
⫺1
2x − 1, if x ≥ 1/2
−(2x − 1), if x < 1/2
0.5
1
1.5
x
2
⫺1 ⫺0.5
⫺1
⫺2
⫺2
⫺2
⫺3
⫺3
⫺3
0.5
1
1.5
2
x
Examples and Explorations
EXAMPLE 1
Finding domains of power functions
Find the domains of the following power functions:
(a) f (x) = 3x −2
1
2
(b) g(x) = x 3/4 (c) h(x) = 2x 4/5
(d) k(x) = 8x −1/2
SOLUTION
(a) If we rewrite the power function so that any roots or denominators are visible, the
3
calculation of the domain is obvious: f (x) = 3x −2 = 2 is defined everywhere but at
x
x = 0, so its domain is (−∞, 0) ∪ (0, ∞).
1
(b) The function g(x) = x 3/4 =
2
is [0, ∞).
√
( 4 x )3
is defined everywhere except x < 0, so its domain
2
2
is defined for all x = 0, so its domain is (−∞, 0) ∪
( x )4
(c) The function h(x) = 2x −4/5 = √5
(0, ∞).
8
x
(d) The function k(x) = 8x −1/2 = √ fails to be defined for x = 0 and for x < 0, so its
domain is (0, ∞).
EXAMPLE 2
Modeling a graph with a polynomial
Explain why the graph shown here could be modeled with a polynomial function f . Then say
what you can about the degree and leading coefficient of f , and find a possible equation for f (x).
y
9
6
3
⫺4 ⫺3 ⫺2 ⫺1
⫺3
1
2
3
4
x
⫺6
⫺9
CONFIRMING PAGES
11:11
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.3
41
Algebraic Functions
SOLUTION
This graph is defined on all of R, is smooth and unbroken everywhere, and has no asymptotes, so it could be part of the graph of a polynomial function f . We will now find a polynomial whose graph is like the one pictured. Since both ends of the graph point upwards,
we know that the degree of f must be even and the leading coefficient of f must be positive.
The fact that the graph has three roots means that degree of f is at least 3. The fact that the
graph has three local extrema means that the degree of f is in fact at least 4.
The roots of f are x = −3, x = −1, and x = 2. Note that the function f behaves differently at the root x = 2 than at the other roots. Near x = 2 the graph appears to have a
quadratic type of shape. This means that x = 2 is a repeated root of f and therefore that
(x−2) is a factor of f more than once. Given all of this information, we see that one possible
form for f (x) is
f (x) = A(x + 3)(x + 1)(x − 2)2 ,
where A is some positive constant. The y-intercept (0, 6) marked on the graph can help us
determine A. Since f (0) = 6, we have
6 = A(0 + 3)(0 + 1)(0 − 2)2
=⇒
6 = A(3)(1)(4)
=⇒
6 = 12A
=⇒
1
2
A= .
Therefore a function that could have the given graph is the quartic polynomial function
1
2
f (x) = (x + 3)(x − 1)(x − 2)2 .
EXAMPLE 3
Making a rough graph of a rational function
Without a calculator, sketch a rough graph of the function f (x) =
2(x − 1)2 (x + 2)
.
(x − 1)(x + 1)(x − 2)
SOLUTION
By Theorems 0.17 and 0.18, we can see immediately from the factors in the numerator and
denominator that the graph of f will have
a hole at x = 1;
a root at x = −2;
vertical asymptotes at x = −1 and x = 2;
a horizontal asymptote at y =
an
bn
=
1
2
= 2.
A quick sign analysis tells us where the graph of f is above or below the x-axis. We must
check the sign of f on each subinterval between the roots and non-domain points:
2
DNE
1
DNE 1
2
f
Note that on this number line we include tick-marks only at the locations where f (x) is
zero or does not exist. The unlabeled tick-marks are the locations where f (x) is zero, and
the ones labeled “DNE” are the locations where f (x) is not defined.
Plotting a few key points will help us make a more accurate graph. The y-intercept of
the graph is f (0) = 2. The height of the hole at x = 1 will be the value of f at x = 1 after
cancelling common factors:
2(1 − 1)(1 + 2)
2(0)(3)
=
= 0.
(1 + 1)(1 − 2)
(2)(−1)
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
42
Chapter 0
November 26, 2012
Functions and Precalculus
Finally, just to get a value to the right side of the graph, we also calculate f (3) = 5. The
figure that follows at the left shows some of the information we have collected. The figure
at the right uses this information and the sign analysis to fill in a rough sketch of the graph.
Some information about the graph
A rough sketch of f
y
y
5
5
2
2
2 1
1 2 3
x
2 1
1 2 3
x
EXAMPLE 4
Writing an absolute value expression as a piecewise-defined function
Write the function f (x) = |x 2 − 1| as a piecewise-defined function, and use this piecewisedefined function to calculate f (−2), f (0), and f (1). Then sketch a graph of f .
SOLUTION
When x 2 − 1 is positive or zero, its absolute value will remain x 2 − 1. When x 2 − 1 is negative, its absolute value will be −(x 2 − 1). Therefore we have
x 2 − 1, for all x with x 2 − 1 ≥ 0
f (x) = |x 2 − 1| =
−(x 2 − 1), for all x with x 2 − 1 < 0.
Although this is one way to write |x 2 − 1| as a piecewise-defined function, it is difficult to
work with. For example, to evaluate f (−2) we would need to know whether x 2 − 1 ≥ 0 or
x 2 − 1 ≤ 0 when x = −2. To simplify this piecewise-defined function, we need to rewrite
the conditions as intervals of x-values. Since x 2 − 1 ≥ 0 when x ≥ 1 or x ≤ −1, and
x 2 − 1 < 0 when −1 < x < 1, we have
⎧
⎨ x 2 − 1, if x ≤ −1
2
f (x) = | x − 1 | = −(x 2 − 1), if − 1 < x < 1
⎩
x 2 − 1, if x ≥ 1.
Now we can easily use our formula to calculate f when x = −2, x = 0, and x = 1:
f (−2) = (−2)2 − 1 = 3
2
f (0) = −((0) − 1) = 1
2
f (1) = (1) − 1 = 0
(use first case, since −2 ≤ −1)
(use middle case, since −1 < 0 < 1)
(use last case, since 1 ≥ 1).
Of course, we could have just substituted our x-values into the expression |x 2 − 1|, but
we are practicing writing absolute value functions as piecewise-defined functions. This is
a skill that will come in handy when we later try to differentiate or integrate functions that
involve absolute values.
The graph of y = |x 2 − 1| is the same as the graph of y = x 2 − 1 on the intervals
(−∞, −1] and [1, ∞). On the interval (−1, 1), the quantity |x 2 − 1| has the sign opposite that of x 2 − 1, so its graph is the reflection of the negative parts of x 2 − 1 across the
x-axis, as the following figure shows:
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.3
Algebraic Functions
43
y
4
3
2
1
3
2
1
1
2
3
x
1
TEST YOUR
? UNDERSTANDING
Why is not every power function also a polynomial function?
Why is every polynomial function also a rational function?
Given that the degree of a polynomial function is the integer n that represents the
highest power of x with a nonzero coefficient in the polynomial, why do you think we
say that the degree of the zero polynomial (the function f defined by f (x) = 0 for all
values of x) is undefined?
Calculators are notoriously bad at graphing rational functions. Sometimes they connect
a graph over its vertical asymptotes, and most times the holes of a rational function are
not immediately clear from a calculator graph. What do you think causes calculators to
make these errors?
In Example 4, why does the function f (x) = |x 2 − 1| have break points at x = −1 and
x = 1?
EXERCISES 0.3
Thinking Back
Basic algebra review: The following exercises will help you
review your skill with exponents, fractions, and factoring.
Calculate the value of
8
27
−2/3
by hand.
Factoring after root-guessing: Factor as much as possible each
of the polynomial and rational functions that follow. In these
exercises it is necessary to guess a root and use synthetic division to get started with factoring.
x −3 − x −2
in the form Ax k for some real numx −1 − 1
bers A and k.
f (x) = 2x 4 + 6x 2 − 8
Factor 16x 6 − 81x 2 as much as possible.
f (x) =
Solve the equation
Write
x 3 + x 2 − 2x
= 0.
x 2 − 4x + 3
1
= 2x − 1.
Solve the equation 2
3x − 2x − 1
f (x) = x 3 + 4x 2 − 11x + 6
x 3 + 6x 2 + 3x − 10
x 5 + 3x 4 − x − 3
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The sum of any two algebraic functions
is itself an algebraic function.
(b) True or False: Every power function is a polynomial
function.
(c) True or False: Every polynomial function is a rational
function.
(d) True or False: For any real numbers a and b, the polynomial function f (x) = x 4 − ax 3 + bx + 7 has exactly
four real roots.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
44
Chapter 0
November 26, 2012
Functions and Precalculus
(e) True or False: If f is a polynomial function with k real
roots, then the degree of f must be at most k.
(f) True or False: If f is a polynomial function with k
turning points, then the degree of f must be at least
k + 1.
(g) True or False: A rational function is a quotient of two
polynomials, where the denominator polynomial is
not the zero polynomial.
(h) True or False: Every rational function has a horizontal
asymptote.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Three functions that are algebraic.
(b) Three functions that are algebraic and involve quotients, but are not rational.
(c) Three functions that fail to be algebraic.
3. Sketch graphs of y = −2x 3 and y = 3x −2 by hand, without using a calculator. Label three points on each graph.
4. Sketch graphs of y = −3x1/2 and y = x 2/3 by hand, without using a calculator. Label three points on each graph.
5. Sketch graphs of y = x 2 , y = x 4 , and y = x 6 by hand,
without a calculator, on the same set of axes. Label three
points on each graph.
6. Sketch graphs of y = x 3 , y = x 5 , and y = x 7 by hand,
without a calculator, on the same set of axes. Label three
points on each graph.
7. If f (x) = Ax k is a power function, is its reciprocal
13. Explain why the graph of f (x) =
have a hole in it at x = −2.
14. Use a calculator to graph the function f (x) =
15. Construct an equation of a rational function whose graph
has no roots, no holes, vertical asymptotes at x = ±2, and
a horizontal asymptote at y = 3.
16. Construct an equation of a rational function whose graph
has a hole at the coordinates (−2, 0), vertical asymptotes
at x = 1 and x = −3, and a horizontal asymptote at y =
−5.
17. Suppose f is a rational function with roots at x = 1 and
x = 3, a hole at x = −1, a vertical asymptote at x = 2, and
a horizontal asymptote at y = −1.
(a) Sketch three possible graphs of f . Make the graphs
as different as you can while still having the given
characteristics.
(b) Write down the equations of three functions f that
have the given properties. (Your equations do not
have to match your graphs from part (a).)
18. Suppose f is a rational function with root at x = −2, holes
at x = 0 and x = 3, a vertical asymptote at x = 1, and no
horizontal asymptote.
(a) Sketch three possible graphs of f . Make the graphs
as different as you can while still having the given
characteristics.
(b) Write down the equations of three functions f that
have the given properties. (Your equations do not
have to match your graphs from part (a).)
a power function? If so, write the reciprocal in the form
Cx r for some real numbers C and r.
8. Suppose f (x) = Ax k is a power function, where k is an
integer.
For each graph y = f (x) shown, sketch the graph of y = | f (x)|.
y
19.
9. For the polynomial f (x) = x(3x + 1)(x − 2)2 , determine
the leading coefficient, leading term, degree, constant
term, and coefficients a1 and a3 .
10. For the polynomial f (x) = x 5 (1 − 2x)(1 − x 3 ), determine
the leading coefficient, leading term, degree, constant
term, and coefficients a1 and a5 .
11. Use the quadratic formula to explain why a quadratic
polynomial function f (x) = ax 2 + bx + c is irreducible if,
and only if, the discriminant b2 − 4ac is negative. Then
use the discriminant to show that f (x) = 3x 2 + 2x + 6 is
irreducible.
12. Give an example of each of the following types of polynomials:
(a) A quintic polynomial with only one real root.
(b) A polynomial, all of whose roots are rational, but
non-integer, numbers.
(c) A polynomial with integer coefficients that has four
real roots, only two of which are integers.
x2 + x − 2
.
x2 − 4
This graph has a hole, where the function is not defined;
determine the location of the hole, and trace along the
graph on your calculator until you find it there.
1
also
f (x)
(a) What must be true about k for f to be one-to-one,
and why?
(b) If f is one-to-one, is its inverse f −1 also a power function? If so, write the inverse in the form f −1 (x) = Cx r
for some real numbers C and r.
(x − 1)(x + 2)
does not
(x + 2)2
y
20.
3
3
2
2
1
1
3 2 1
1
1
2
3
x
2
2
3
3
y
21.
3
3 2 1
1
15
10
2
5
1
3
6
x
2
3
1
2
3
x
y
22.
3
5
1
3 2 1
1
10
2
15
3
CONFIRMING PAGES
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
0.3
23. An alternative
definition for the absolute value function
√
is |x| = x 2 . Explain in your own words why this definition is equivalent to that of the absolute value function given in Definition 0.19. This exercise shows that the
absolute value function is algebraic, not just piecewise
algebraic.
45
Algebraic Functions
24. Suppose that f is a function defined in pieces, using some
functions g and h, as follows:
g(x), if x ≤ 2
f (x) =
h(x), if x > 2.
Will the piece of the graph on (−∞, 2] always “match up”
with the piece of the graph on (2, ∞)? Why or why not?
Use graphs to illustrate your answer.
Skills
Find the domain and the zeroes of each function in Exercises 25–34. Check your answers afterwards with graphs.
1
4
−4
25. f (x) = − x 2/3
26. f (x) = √
6
27. f (x) = 2x −7/4
28. f (x) =
29. f (x) = (x 2 − 1)−1/4
30. f (x) = |x 2 − 9|−3/4
31. f (x) = √
3
√
32. f (x) = 4 x −2 + 1
33. f (x) =
√
3−x
x 2 − 3x − 4
x2 − 1
x 3 − 7x + 6
x 7/3
34. f (x) =
x 3 + x 2 − 2x
2x 3 − x 2 − 6x
3x −3
x −1/4 − x 3/4
Find equations for each of the functions described in Exercises 35–42.
51. f (x) =
2x 3 + 4x 2 − 6x
x2 − 4
53. f (x) =
(x + 1)(x − 3)(x + 2)
(x + 2)(x − 3)
54. f (x) =
(x − 1)(x + 2)
(x + 1)(x − 1)
2x 3 + 3x 2 − 2x − 3
x 2 − 2x − 3
2
x (x − 3)(x + 2) 56. f (x) = x−1
Given that the graph of f (x) = sin x is as shown here, sketch
graphs of each of the following absolute value transformations g(x):
y
2
1
37. The linear function whose graph is parallel to y = 2x + 1
and passes through the point (−1, 4).
38. A power function whose graph passes through (0, 0) and
(1, 3).
41. A polynomial function whose graph passes through (0, 0),
(2, 0), and (4, 0).
42. A polynomial function whose graph passes through (0, 0),
(2, 0), (4, 0), and (1, 2).
Sketch rough graphs of the functions in Exercises 43–56
without using a calculator or graphing utility. Be as accurate
as you can, and identify any roots, holes, or asymptotes.
√
44. f (x) = (x + 3)2/3 − 2
43. f (x) = 3 x + 1
45. f (x) =
1
16 − x 4
46. f (x) = x 4 − 6x 2 + 9
47. f (x) = 2x 4 − x 3 − x 2
48. f (x) = x 4 − 2x 2 + 1
49. f (x) = x 3 − 2x 2 − 4x + 8
50. f (x) = x(x + 2)(x − 3)2
(x 2 − 4)2
2x 2 − 3x − 2
55. f (x) =
35. The linear function whose graph has slope −1 and passes
through the point (3, −2).
36. The linear function whose graph passes through the
points (−2, 1) and (3, −4).
39. A power function whose graph passes through (0, 0) and
(2, 8).
40. A polynomial function whose graph passes through
(−2, 0), (1, 0), and (3, 0).
52. f (x) =
⫺3π ⫺2π ⫺π
π 2π 3π
x
⫺1
⫺2
57. g(x) = | sin x|
58. g(x) = sin |x|
59. g(x) = −2 + | sin x|
60. g(x) = |− 2 + sin x|
For each graph in Exercises 61–72, find a function whose
graph looks like the one shown. When you are finished, use
a graphing utility to check that your function f has the properties and features of the given graph.
y
61.
y
62.
1
x
⫺2
⫺1
CONFIRMING PAGES
⫺2
x
11:11
TKmaster2010
WHF00153/FREE087-Taalman
46
Chapter 0
Functions and Precalculus
y
63.
November 26, 2012
y
64.
y
69.
y
70.
4
4
1
3 2
2
x
1
2
y
65.
1
1
x
1
y
66.
y
71.
3
x
1
y
72.
2
18
6
x
1 2
8
1
6
2
2
y
x
1
y
68.
1
2
3
x
2
67.
1
2
4
x
1
3
x
1
x
1
Write each function in Exercises 73–80 as a piecewise-defined
function where each piece is defined on an interval of
x-values. Then sketch a labeled graph of each piecewisedefined function.
3
5.67
1
x
2 3
73. f (x) = |5 − 3x|
74. f (x) = |1.5x + 2.3|
75. f (x) = |x + 1|
76. f (x) = |x 2 − 4|
77. f (x) = |9 − x 2 |
78. f (x) = |3 − 4x + x 2 |
79. f (x) = |x 2 − 3x − 4|
80. f (x) = |x 3 + x 2 − 2x|
2
Applications
The following table for Exercises 81 and 82 describes the
number of cars that were on a particular 1-mile stretch of
Route 97 at t hours after 6:00 a.m. Monday morning:
t (hours after 6:00 a.m.)
0
1
2
3
N (cars on 1 mile of road)
0
28
12
18
For Exercises 83 and 84, suppose that Emmy is investigating a release of toxins from a tank farm on the Hanford
nuclear reservation into groundwater. The groundwater eventually forms a spring that runs into the Columbia River. The
following table describes the amount of toxins in the spring:
Years after 2006
Concentration of toxin (ppm)
81. Make a plot of the data given in the preceding table. Suppose you wanted to find a polynomial function N(t) that
passes through each of the four data points. What is the
minimum degree that this polynomial function could be,
and why?
82. Given the same table of car densities,
(a) Find a polynomial function N(t) that goes through all
four data points in the table. (Hint: Use the data to solve
for the coefficients.)
(b) Overlay a plot of the function N(t) you found with the
plot of the data. Does N(t) look the way you expected
it to?
(c) Use your function N(t) to predict the number of cars
that were on that 1-mile stretch at 7:30 a.m. How
many cars does your function predict will be there at
3 p.m.? On what time interval does your model function make practical sense?
0
1
2
0.53
0.65
0.74
83. Emmy believes that the leak is not getting larger, so that
the concentration of toxin in the spring will approach
some steady constant value. She wants to make estimates
of the date on which the leak started and of the eventual
steady concentration of toxin. Is a polynomial the best
choice of a function to fit to the data for this purpose?
Why or why not?
84. Emmy wants to fit a rational function of the form
T(t) =
at + b
to the data, where t is the number of years
t+d
after 2006.
(a) What values should Emmy use for the coefficients a, b
and d?
(b) Use your rational function model to estimate the date
that the leak started.
(c) What asymptotic value will the concentration approach as time t increases without bound?
(d) What are the potential problems in using this model?
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
47
Exponential and Trigonometric Functions
Proofs
85. Prove that the composition of two power functions is a
power function.
86. Prove that the product of two power functions is a power
function.
90. Prove that the sum or product of two rational functions is
a rational function.
87. Prove that the composition of two linear functions is also
a linear function.
88. Prove that the sum of two polynomial functions is a
polynomial.
92. Prove that the domain of a rational function f (x) =
89. Prove that the product of two cubic polynomials is a
polynomial of degree six.
91. Prove that (a) every constant function is linear and
(b) every linear function is a polynomial.
the set { x | q(x) = 0 }.
93. Prove that the graph of a rational function f (x) =
a root at x = c if and only if p(c) = 0 but q(c) = 0.
p(x)
is
q(x)
p(x)
has
q(x)
Thinking Forward
Algebra for derivatives: Simplify and rewrite each expression
until you can cancel the h in the denominator:
(x + h)3 − x 3
h
(x + h) 1/2 − x 1/2
h
(x + h)−2 − (x)−2
h
(x + h)−1/2 − x−1/2
h
0.4
Alternative algebra for derivatives: Simplify and rewrite each
expression until you can cancel a common factor from the
numerator and denominator:
t3 − x3
t−x
t 1/2 − x 1/2
t−x
t−2 − x −2
t−x
t−1/2 − x−1/2
t−x
EXPONENTIAL AND TRIGONOMETRIC FUNCTIONS
Definitions and properties of exponential and logarithmic functions
Definitions and properties of trigonometric and inverse trigonometric functions
Graphs and equations involving transcendental functions
Exponential Functions
Functions that are not algebraic are called transcendental functions. In this book we will
investigate four basic types of transcendental functions: exponential, logarithmic, trigonometric, and inverse trigonometric functions. Exponential functions are similar to power
functions, but with the roles of constant and variable reversed in the base and exponent:
DEFINITION 0.21
Exponential Functions
An exponential function is a function that can be written in the form
f (x) = Ab x
for some real numbers A and b such that A = 0, b > 0, and b = 1.
There is an important technical problem with this definition: We know what it means to
raise a number to a rational power by using integer roots and powers, but we don’t know
what it means to raise a number to an irrational power. We need to be able to raise numbers
to irrational powers to talk about exponential functions; for example, if f (x) = 2 x , then we
need to be able to compute f (π) = 2 π . One way to think of b x where x is irrational is as a
limit:
b x = lim b r .
r→x
r rational
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
48
Chapter 0
November 26, 2012
Functions and Precalculus
The “lim” notation will be explored more in Chapter 1. For now you can just imagine that
if x is rational, then we can approximate b x by looking at quantities b r for various rational
numbers r that get closer and closer to the irrational number x. For example, 2 π can be
approximated by 2 r for rational numbers r that are close to π :
100
2 π ≈ 23.14 = 2314/100 =
2314 .
As we consider rational numbers r that are closer and closer to π, the expression 2 r will get
closer and closer to 2 π ; see Exercise 4. In Chapter 7 we will give a more rigorous definition
of exponential functions as the inverses of certain accumulation integrals.
We will assume that you are familiar with the basic algebraic rules of exponents, for
example that b x+y = b x b y , that b0 = 1 for any nonzero b, and that (b x ) y = b xy . Proving
those rules requires the more rigorous definition of exponential functions that we will see
in Chapter 5, so for the moment we will take these algebraic rules as given. From those
basic rules it follows that an exponential function f (x) = b x is one-to-one, and that b x is
never zero for any value of x. (See Exercises 85 and 86.)
Interestingly, the most natural base b to use for an exponential function isn’t a simple
integer, like b = 2 or b = 3. Instead, for reasons that will become clear when we study
derivatives, the most natural base is the irrational number known as e, and the function e x
is therefore called the natural exponential function. An approximation of the number e
to 65 digits is:
2.7182818284590452353602874713526624977572470936999595749669676277 . . . .
Of course, since e is an irrational number, we cannot define it just by writing an approximation of e in decimal notation; we will define e properly once we cover limits in
Chapter 1.
In Exercise 88 you will prove that every exponential function can be written so that its
base is the natural number e, as the next theorem states:
THEOREM 0.22
Natural Exponential Functions
Every exponential function can be written in the form
f (x) = Ae k x
for some real number A and some nonzero real number k.
Every exponential function has a graph similar to either the exponential growth graph
that follows at the left or the exponential decay graph at the right, depending on the values
of k and b. Of course, if the coefficient A is negative, then the graph of f (x) = Ae k x or
f (x) = Ab x will be a reflection of one of these two graphs over the x-axis.
f (x) = e k x with k < 0,
f (x) = b x with 0 < b < 1
f (x) = e k x with k > 0,
f (x) = b x with b > 1
y
2
1
y
4
4
3
3
2
2
1
1
1
2
x
2
1
1
CONFIRMING PAGES
2
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
Exponential and Trigonometric Functions
49
Logarithmic Functions
Since every exponential function b x is one-to-one, every exponential function has an inverse. These inverses are what we call the logarithmic functions:
DEFINITION 0.23
Logarithmic Functions as Inverses of Exponential Functions
The inverse of the exponential function b x is the logarithmic function
logb x.
As a special case, the inverse of the natural exponential function e x is the natural logarithmic function
loge x = ln x.
We require that the base b satisfy b > 0 and b = 1, because these are exactly the conditions we must have for y = b x to be an exponential function. In Section 7.7 we will define
logarithms another way, in terms of integrals and accumulation functions.
You should already be familiar with the algebraic rules of logarithms, but we restate
them here in case you need a refresher; see Exercises 90–94 for proofs.
THEOREM 0.24
Algebraic Rules for Logarithmic Functions
For all values of x, y, b, and a for which these expressions are defined, we have
(a) logb x = y if and only if b y = x
(b) logb (b x ) = x
(c) blogb x = x
(d) logb (x a ) = a logb x
(e) logb (xy) = logb x + logb y
(f) logb
(g) logb
1
x
= − logb x
x
y
= logb x − logb y
(h) logb x =
log a x
log a b
The first three properties follow from properties of inverse functions, and tell us that logb x
is the exponent to which you have to raise b in order to get x. For example, log2 8 is the
power to which you have to raise 2 to get 8; since 23 = 8, we have log2 8 = 3. All of these
rules also apply to the natural exponential function, because ln x is just logb x with base
b = e.
Properties (d) and (e) follow from the algebraic rules of exponents, and properties (f)
and (g) are their immediate consequences. The final property in Theorem 0.24 is called the
base conversion formula, because it allows us to translate from one logarithmic base to
another. The base conversion formula is especially helpful for converting to base e or base
ln2
10 so that we can calculate logarithms on a calculator. For example, log7 2 is equal to
,
ln7
which we can approximate using the built-in “ln” key on a calculator.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
50
Chapter 0
November 26, 2012
Functions and Precalculus
The graphs of logarithmic functions can be obtained easily from the graphs of exponential functions by reflection over the line y = x, resulting in the following graphs:
y = b x and y = logb x with b > 1
y = b x and y = logb x with 0 < b < 1
y
y
5
5
4
4
3
3
2
2
1
1
2 1
1
1
2
3
4
5
x
2 1
1
1
2
3
4
5
x
2
2
Trigonometric Functions
There are six trigonometric functions defined as ratios of side lengths of right triangles, or, more generally, as ratios of coordinate lengths on the unit circle. We now provide a quick review of the definitions of these functions and their graphical and algebraic
properties.
We can place any angle in a standard position in the xy-plane by placing its vertex at
the origin and its initial edge along the positive side of the x-axis, as shown next at the
left. The angle then opens up in either a counterclockwise or clockwise direction until it
reaches its terminal edge. A positive angle is measured counterclockwise from its initial
edge, while a negative angle is measured clockwise from its initial edge.
terminal
edge
y
y
y
2π
its
un
θ
initial
edge
x
1 unit
1 unit
1 radian
x
x
Now consider the unit circle shown in the center diagram. Since the unit circle has radius r = 1 unit, its circumference is C = 2πr = 2π(1) = 2π units. That’s a circumference
of approximately 6.283185 units, which certainly is not as nice as a number like 360 that
we can easily divide into integer-sized pieces. However, we can still measure everything in
terms of this circumference by defining a new unit of angle measure called a radian that
represents the size of an angle in standard position whose terminal edge intersects the unit
circle after an arc length of 1 unit, as shown in the diagram at the right.
Since the distance all the way around the circle is 2π units, the distance halfway around
π
is π units and the distance one-quarter of the way around the circle is units. This means
that an angle of 90◦ measures
π
2
2
radians, an angle of 180◦ measures π radians, and an
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
51
Exponential and Trigonometric Functions
angle of 360◦ measures, of course, 2π radians. The following three diagrams illustrate some
common positive angles in radian measure around the unit circle:
y
y
π
2
π
y
3π
4
π
4
x
0
π
3
π
6
5π
6
x
5π
4
3π
2
2π
3
x
7π
6
7π
4
4π
3
11π
6
5π
3
Of course we can also consider negative angles; for example, the angle that opens up in
the clockwise direction for one quarter of the distance around the bottom half of the circle
π
has radian measure − . Its terminal edge intersects the unit circle in the same location as
the angle
7π
4
4
shown in the middle figure. We can also consider angles that go more than
once around the circle; for example, the angle
same point as the angle
π
2
5π
2
= 2π +
π
2
intersects the unit circle at the
in the diagram at the left.
Given any angle θ in standard position, the terminal edge of θ intersects the unit circle
at some point (x, y) in the xy-plane. We will define the height y of that point to be the sine
of θ, while the cosine of θ will be defined as the x-coordinate of that point.
DEFINITION 0.25
Trigonometric Functions for Any Angle
Given any angle θ measured in radians and in standard position, let (x, y) be the point
where the terminal edge of θ intersects the unit circle. The six trigonometric functions
of an angle θ are the six possible ratios of the coordinates x and y for θ :
y
sin θ = y
θ
x
(x, y)
(cos θ, sin θ)
csc θ =
1
y
cos θ = x
tan θ =
y
x
1
x
cot θ =
x
y
sec θ =
Notice that the sine and cosine functions determine the remaining four trigonometric
sinθ
, and the last three trigonometric functions are the
functions, since tan θ is the ratio
cosθ
reciprocals of the first three.
You should already be familiar with the basic trigonometric identities, but they are repeated next for your review; see Exercises 95–100 for proofs. The first Pythagorean identity,
the even–odd identities, and the shift identities follow easily from the definitions of the
trigonometric functions. The sum identities follow from a geometric argument that we will
not get into here. The remaining identities can all be proved from the previous identities.
In the following identities we are using the notation sin2 x as shorthand for (sin x)2 .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
52
Chapter 0
THEOREM 0.26
November 26, 2012
Functions and Precalculus
Basic Trigonometric Identities
Pythagorean Identities
2
Even–Odd Identities
2
sin θ + cos θ = 1
Shift Identities
π
= sin θ
cos θ −
2
π
= cos θ
sin θ +
sin(−θ) = − sin θ
tan2 θ + 1 = sec2 θ
cos(−θ) = cos θ
1 + cot 2 θ = csc2 θ
tan(−θ ) = − tan θ
2
sin(θ + 2π) = sin θ
cos(θ + 2π) = cos θ
Sum Identities
Difference Identities
sin (α + β) = sin α cos β + sin β cos α
sin (α − β) = sin α cos β − sin β cos α
cos (α + β) = cos α cos β − sin α sin β
cos (α − β) = cos α cos β + sin α sin β
Double-Angle Identities
Alternative Forms
Alternative Forms
1 − cos 2θ
2
1 + cos 2θ
2
cos θ =
2
2
sin2 θ =
sin 2θ = 2 sin θ cos θ
cos 2θ = 1 − 2 sin θ
cos 2θ = cos2 θ − sin2 θ
cos 2θ = 2 cos2 θ − 1
The graphs of the six trigonometric functions are shown next. Each of the graphs in the
second row is the reciprocal of the graph immediately above it. Remember that you can use
1
the graph of a function f to sketch the graph of its reciprocal . In particular, the zeros of f
f
1
f
will be vertical asymptotes of , large heights on the graph of f will become small heights
1
f
on the graph of , and vice versa.
y = sin x
y = cos x
y = tan x
y
y
y
3
2
2
2
1
1
3π 2π π
π 2π 3π
x
1
3π 2π π
1
π
2π 3π
x
1
3π 2π π
1
2
y = sec x
y
y = cot x
y
y
3
3
2
2
2
1
1
1
π
2π 3π
x
x
3
3
3π 2π π
1
2π 3π
2
2
y = csc x
π
3π 2π π
1
π
2π 3π
x
3π 2π π
1
2
2
2
3
3
3
CONFIRMING PAGES
π
2π 3π
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
53
Exponential and Trigonometric Functions
Inverse Trigonometric Functions
None of the six trigonometric functions are one-to-one, but after restricting domains we
can construct the so-called inverse trigonometric functions. In this section we will focus
on the inverses of only three of the six inverse trigonometric functions: Those for sine,
tangent, and secant. (The inverses of these functions will be particularly useful to us in
Chapter 5 when we study integration techniques, and the inverses of the remaining three
trigonometric functions would add no more to that discussion.) There are many different
restricted domains that we could use to obtain partial inverses to these three functions. We
need to pick one restricted domain for each function and stick with it. In this text we will
use the restricted domains shown below.
y = sin x restricted
π πto
the domain − ,
y = tan x restricted
π πto
the domain − ,
y = secx restricted
πto π
the domain 0,
∪
,π
y
y
y
2 2
2 2
2
2
1
1
π
π
π
π
2
2
x
π
π
2
1
π
2
1
π
x
π
π
2
1
π
2
π
x
1
Each of these restricted functions is one-to-one and thus invertible. The inverses of these
restricted functions, respectively, are the inverse sine, inverse tangent, and inverse secant
functions.
The Inverse Trigonometric Functions
−1
(a) The inverse
sine
function sin x is the inverse of the restriction of sin x to the
π π
interval − , .
DEFINITION 0.27
2
2
−1
(b) The inverse
tangent
function tan x is the inverse of the restriction of tan x to the
π π
2 2
interval − ,
.
(c) The inverse
sec−1 x is the inverse of the restriction of sec x to the
secant
function
interval 0,
π
2
∪
π
,π
2
.
Notice that since the inputs to the trigonometric functions are angles, it is the outputs of the
inverse trigonometric functions that are angles. We will interchangeably use the alternative
notations arcsin x, arctan x, and arcsec x for these inverse trigonometric functions.
CAUTION
Although we use the notation sin2 x to represent (sin x)2 and the notation x −1 to represent
1
1
, the notation sin−1 x does not represent
. Inverse functions in general have nothing
x
sinx
to do with reciprocals, despite what one might imagine from the notation.
All of the properties of sin−1 x, tan−1 x, and sec−1 x come from the fact that they are
the inverses of the restricted functions sin x, tan x, and sec x. For example, we can graph the
inverse trigonometric functions simply by reflecting the graphs of the restricted trigonometric functions over the line y = x, as follows:
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
54
Chapter 0
November 26, 2012
Functions and Precalculus
y = sin−1 x
y = tan−1 x
y = sec−1 x
y
y
y
π
π
π
π
2
π
2
π
2
1
1
x
1
π
x
1
1
π
x
1
π
2
2
2
π
π
π
Although sin−1 x and (restricted) sin x are transcendental functions, their composition
sin (sin x) = x is algebraic. This is obvious because these functions are inverses of each
other. However, something more general and surprising is true: The composition of any
inverse trigonometric function with any trigonometric function is algebraic; see Example 4.
−1
Examples and Explorations
EXAMPLE 1
Finding values of transcendental functions by hand
Calculate each of the following by hand, without a calculator:
(a) log6 3 + log6 12
(b) cos
5π
6
(c) sin−1
1
2
SOLUTION
(a) log6 3 is the exponent to which we would have to raise 6 to get 3; think 6? = 3. It is not
immediately apparent what this exponent is. Similarly, it is not clear how to calculate
log6 12 without a calculator. However, using the additive property of logarithms we
can write
log6 3 + log6 12 = log6 (3 · 12) = log6 36 = 2.
The final equality holds because 62 = 36.
5π
lies on the unit circle. If
(b) The diagram that follows at the left shows where the angle
6
we draw a line from the point (x, y) where the angle meets the unit circle to the x-axis,
we obtain a triangle whose reference angle is 30◦ . Using the known side lengths of
a 30–60–90 triangle with hypotenuse of length 1, we can label the side lengths of our
reference triangle, as shown
in the middle figure. This in turn means that we know the
√
3 1
coordinates (x, y) = − ,
of the point at which the terminal edge of θ intersects
2
2
the unit circle. Therefore cos
5π
Angle θ =
has
6
5π
6
=−
√
3
.
2
Side lengths of a 30–60–90
triangle with hypotenuse 1
reference angle 30◦
y
y
5π
6
θ
30°
π
π π
is the angle in − ,
6
2 2
1
whose sine is equal to
2
y
( 23 , 21 )
5π
6
x
1
2
1
30°
θ
π
6
1
2
5π
6
x
3
2
CONFIRMING PAGES
30°
x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
55
Exponential and Trigonometric Functions
1
1
(c) If θ = sin−1 , then we must have sin θ = . There are infinitely many angles whose
2
2
1
π π
sine is , but only one of those angles is in the restricted domain − , . Thus θ =
2 2
2
1
π π
1
is the unique angle in − ,
whose sine is , as shown in the figure at
sin−1
2
2
2
2
1
2
the right. Notice that the triangle must be a 30–60–90 triangle (since its height is ),
and therefore the angle θ we are looking for must be 30◦ (i.e.,
1
sin−1
2
EXAMPLE 2
=
π
.
6
π
6
radians). Therefore
Solving equations that involve transcendental functions
Solve each of the following equations:
(a) 3.25(1.72) x = 1000
(c) sec−1 x =
(b) sin θ = cos θ
π
6
SOLUTION
(a) To solve for x we will isolate the expression (1.72) x and then apply the natural logarithm
so that we can get x out of the exponent:
1000
1000
3.25(1.72) x = 1000 =⇒ ln((1.72) x ) = ln
=⇒ x ln(1.72) = ln
.
3.25
3.25
It is now a simple matter to solve for
x=
ln(1000/3.25)
≈ 10.564.
ln(1.72)
(b) If sin θ = cos θ , then θ is an angle whose terminal edge intersects the
√ unit√circle at a
point
(x,
y)
with
x
=
y.
The
only
such
points
on
the
unit
circle
are
(
2/2, 2/2) and
√
√
(− 2/2, − 2/2), as shown in the left-hand diagram that follows. The angles that end
π
at these points are all of the form θ = + π k for some integer k. Thus the solution set
4
3π π 5π 9π
for the equation is . . . , − , , , , . . . .
4
4
4
4
Diagram to solve sec−1 x =
Diagram to solve sin θ = cos θ
y
y
π
4
2
2
2
2
45°
π
6
45°
2
2
2
2
π
6
1
2
30°
x
3
2
x
5π
4
(c) If sec−1 x =
π
,
6
then
x = sec
π
1
1
2
=
π = √3/2 = √3 .
6
cos
6
π
The angle and the reference triangle we used for this calculation are shown in the
6
right-hand diagram.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
56
Chapter 0
EXAMPLE 3
November 27, 2012
Functions and Precalculus
Domains and graphs of transcendental functions
Find the domain of each of the functions that follow. Then use transformations to sketch
careful graphs of each function by hand, without a graphing utility.
(a) f (x) = 5 − 3e1.7x
(b) g(x) =
1
ln(x − 2)
(c) h(x) = 3 sec 2x
SOLUTION
(a) The domain of f (x) = 5−3e 1.7x is R, and the graph of f is a transformation of the exponential growth function e 1.7x shown in the left-hand figure that follows. y = −3e 1.7x
can be obtained by reflecting the leftmost graph over the x-axis and then stretching
vertically by a factor of 3, as shown in the middle figure. The graph of f (x) = 5 − 3e 1.7x
can now be obtained by shifting the middle graph up five units, as shown at the right.
y = e 1.7x
y = −3e 1.7x
y
y
y
6
5
1
5
⫺1
4
⫺1
y = 5 − 3e 1.7x
1
⫺1
x
4
3
3
⫺2
2
2
⫺3
1
1
⫺4
1
⫺1
(b) For the function g(x) =
x
⫺1
⫺5
x
⫺2
⫺6
1
ln(x − 2)
1
⫺1
to be defined at a value x, we must have x − 2 > 0,
and thus x > 2. We must also have ln(x − 2) = 0, which means that x − 2 = 1,
and thus x = 3. Therefore the domain of g(x) is (2, 3) ∪ (3, ∞). To sketch the graph of
1
g(x) =
we start with the graph of y = ln x in the left-hand figure that follows,
ln(x − 2)
translate to the right two units as shown in the middle figure, and then sketch the
reciprocal as shown at the right.
y = ln x
y = ln(x − 2)
y=
y
y
y
3
3
3
2
2
2
1
1
1
1
⫺1
2
3
4
5
x
⫺1
1
2
3
4
5
x
⫺1
⫺2
⫺2
⫺2
⫺3
⫺3
⫺3
(c) The function h(x) = 3 sec 2x =
3
cos2x
of
1
2
3
4
5
x
is defined when cos 2x = 0. The latter condi-
tion occurs when 2x is not an odd multiple of
π
.
4
1
ln(x − 2)
Therefore the domain of h(x) is x =
π
2
and thus when x is not an odd multiple
π
(2k
4
+ 1) for positive integers k. To sketch
the graph of h(x), we start with the graph of y = sec x as follows at the left, stretch vertically by a factor of 3 as shown in the middle figure, and then compress horizontally
CONFIRMING PAGES
22:3
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
57
Exponential and Trigonometric Functions
by a factor of 2 as shown at the right. (See Section 0.2 for a review of transformations
of graphs.)
y = sec x
y = 3 sec x
y
y
EXAMPLE 4
π
2
y
9
9
6
6
6
3
3
3
9
π
y = 3 sec 2x
3
π
π
2
x
π
π
2
3
π
π
2
x
π
π
2
3
6
6
6
9
9
9
π
2
π
x
Simplifying compositions of inverse trigonometric and trigonometric functions
Write cos(sin−1 x) as an algebraic function, that is, a function that involves only arithmetic
operations and rational powers.
SOLUTION
π π
If we define θ = sin−1 x, then sin θ = x and θ must be in the interval − , . Let’s first
2 2
π
consider the case where θ is in the first quadrant 0, ; the reference triangle for such a θ
2
is shown next at the left. If we wish θ to have a sine of x, then the length of the vertical
leg of the triangle must be x. The hypotenuse of the triangle is length 1, since we are on the
unit circle. We could also have considered that the sine of θ is “opposite over hypotenuse”;
thus one triangle involving our angle θ could have an opposite side of length x and a
hypotenuse of length 1. Using the Pythagorean theorem, we find that the length of the
√
remaining leg of the triangle is 1 − x 2 , as shown at the right:
Reference triangle for
π
an angle θ in 0,
Use Pythagorean theorem to
determine length of remaining leg
2
y
θ
1
x
x
θ
1 x2
Now cos θ is the horizontal coordinate of the point on the unit circle corresponding to θ ,
or, in terms of “adjacent over hypotenuse,” we have
√
1 − x2
cos θ =
= 1 − x 2.
1
π
The case where θ is in the fourth quadrant, that is, where θ ∈ − , 0 , is similar and also
2
√
2 . Therefore we have shown that cos(sin−1 x) is equal to the
1
−
x
shows that cos θ =
√
algebraic function 1 − x 2 .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
58
Chapter 0
CHECKING
THE ANSWER
December 10, 2012
Functions and Precalculus
√
To verify the strange fact that cos(sin−1 x) = 1 − x 2 , try evaluating both sides at some
simple x-values. While looking at just a few x-values will not prove that the two expressions
are equal for all x, it will at least give us some evidence that the equality is reasonable.
For example, at x = 0 we have
√
cos(sin−1 0) = cos 0 = 1 and
1 − 02 = 1 = 1,
and at x = 1 we have
cos(sin−1 1) = cos
π
2
= 0 and
1 − 12 =
√
0 = 0.
1
As a less trivial example, consider x = . At this value we have
2
√
−1
cos sin
TEST YOUR
? UNDERSTANDING
π
= cos
6
1
2
3
=
2
and
1−
1 2
=
2
1
1− =
4
√
3
3
=
.
4
2
Why do we require that A = 0 and b > 0, b = 1 in the definition of exponential func-
tions? What would the graphs look like when A = 0, when b < 0, b = 0, or b = 1?
In the reading we calculated log7 2 by finding
same answer if we computed
log10 2
?
log10 7
ln2
ln7
with a calculator. Would we get the
How do you convert from radians to degrees, or vice versa?
How is the graph of the reciprocal of a function related to the graph of that function?
How can that information be useful for remembering the graphs of y = csc x, y = sec x,
and y = cot x?
How are the unit-circle definitions of the trigonometric functions related to the right-
triangle definitions of trigonometric functions?
EXERCISES 0.4
Thinking Back
Algebra with exponents: Write each of the following expressions
in the form Ab x for some real numbers A and b:
(a) 32x+1
(d)
1
2(3 x−4 )
(b) 5 x 23−x
(e)
4(3 x )2
2x
(c) (23x−5 )4
(f)
(1/8) x
3(23x+1 )
Famous triangles, degrees, and radians: The following exercises
will help you review and recall basic trigonometry.
Suppose a right triangle has angles 30◦ , 60◦ , and 90◦
and a hypotenuse of length 1. What are the lengths of
the remaining legs of the triangle?
Suppose a right triangle has angles 45◦ , 45◦ , and 90◦
and a hypotenuse of length 1. What are the lengths of
the remaining legs of the triangle?
What is a radian? Is it larger or smaller than a degree?
Compare an angle of 1 degree with an angle of 1 radian, with both angles in standard position.
Show each of the following angles in standard position on the unit circle, in radians:
4π
17π
3π
(b) −
(c)
(d) 21π
(a)
4
3
6
Inverse functions: Suppose f and g are inverses of each other.
What can you say about f ( g(x)) and g( f (x))?
If f has a horizontal asymptote at y = 0, what can you
say about g?
If f has a y-intercept at y = 1, what can you say about
g?
CONFIRMING PAGES
11:19
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
0.4
Exponential and Trigonometric Functions
59
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The function f (x) = 3e0.5x − 2 is an
exponential function.
(b) True or False: Every exponential function f (x) = Aekx
has a horizontal asymptote at y = 0.
(c) True or False: For all x > 0, ln(x 3 ) = 3 ln x.
log6 x
log2 x
(d) True or False: For all x > 0,
=
.
log2 3
log6 3
(e) True or False: If (x, y) is the point on the unit circle cor7π
responding to the angle − , then x is positive and
3
y is negative.
(f) True or False: The sine of an angle θ is always equal to
the sine of the reference angle for θ.
(g) True or False: For any x, 1 − cos2 (5x 3 ) = sin2 (5x 3 ).
(h) True or False: sec−1 x =
1
.
cos−1 x
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Two exponential functions and their inverses.
(b) Two x-values at which tan x is not defined.
(c) Two x-values at which sec−1 x is not defined.
3. What is the definition of an exponential function, and
how is such a function different from a power function?
Is the function f (x) = x x a power function, an exponential
function, or neither, and why?
4. In this exercise we will examine two ways to think of ab
when b is an irrational number, and in particular, we will
consider what the quantity 2π represents.
(a) One way to define 2π is to think of it as a limit. If
we take a sequence a1 , a2 , a3 , . . . of rational numbers
that approaches π, then the sequence 2a1 , 2a2 , 2a3 , . . .
should approach 2π . Said in terms of limits, this
means that
2π = lim 2a ,
a→π
where each a is assumed to be a rational number. Can
you think of a sequence of rational numbers that gets
closer and closer to π ? (Hint: Think about the decimal
expansion of π .)
(b) Another way to consider 2π is to write it as an infinite
product:
2 π = 2 3 21/10 24/100 21/1000 25/10000 29/100000 · · · .
What will the next term in the product be? How
could 2π equal the product of infinitely many
numbers? Wouldn’t that make 2π infinitely large?
Calculate some of the later terms in the product (for
example, 25/10000 or 29/100000 ), and use these calcula-
tions to argue that even though 2π can be written as a
product of infinitely many numbers, it is not necessarily infinitely large.
√
2r for rational values r
5. Approximate 2 3 by calculating
√
that get closer and√closer to 3. (Hint: You can use the decimal expansion of
√ 3 to get a sequence of rational numbers
that approaches 3.)
6. Why can’t we define the number e just by writing it down
in decimal notation to lots of decimal places?
7. Write the exponential function f (x) = 3e −2x in the form
Ab x for some real numbers A and b. Then write the exponential function g(x) = −2(3x ) in the form Ae k x for some
real numbers A and k.
8. Fill in each blank with an interval of real numbers.
(a) An exponential function f (x) = Ab x represents expoand exponential decay if
nential growth if b ∈
.
b∈
(b) An exponential function f (x) = Ae kx represents exand exponential decay
ponential growth if k ∈
.
if k ∈
(c) Suppose that e kx = b x for some real numbers k and
.
b. Then k ∈ (0, ∞) if and only if b ∈
(d) Suppose that e kx = b x for some real numbers k and
.
b. Then k ∈ (−∞, 0) if and only if b ∈
9. In the definition of the logarithmic function logb x, what
are the allowable values for the base b, and why?
10. Fill in the blanks in each of the following statements.
(a) For all x ∈
, log2 x = y if and only if x =
(b) For all x ∈
, 3log 3 x =
(c) For all x ∈
, log4 (4 x ) =
.
.
.
(d) log2 3 is the exponent to which you have to raise
to get
.
11. The graphs of y = log2 x and y = log 4 x are shown here.
Determine which graph is which, without using a calculator. (Hint: Think about the graphs y = 2 x and y = 4 x , and
then reflect those graphs over the line y = x.)
y = log2 x and y = log 4 x
y
2
1
1
2
3
4
x
⫺1
⫺2
12. State the algebraic properties of the natural logarithm
function that correspond to the eight properties of logarithmic functions in Theorem 0.24.
CONFIRMING PAGES
11:28
TKmaster2010
WHF00153/FREE087-Taalman
60
Chapter 0
November 26, 2012
Functions and Precalculus
13. Use algebraic properties of logarithms, the graph of y =
ln x, and your knowledge of transformations
to sketch
1
.
graphs of f (x) = ln(x 2 ) and g(x) = ln
x
x+1
≥ 0.
14. Solve the inequality ln
x−1
15. Give a mathematical definition of sin θ for any angle θ .
Your definition should include the words “unit circle,”
“standard position,” “terminal,” and “coordinate.”
16. Give a mathematical definition of tan θ for any angle θ .
Your definition should include the words “unit circle,”
“standard position,” “terminal,” and “coordinate.”
17. Use the
why
definition of the sine
function to explain
π
9π
7π
is equal to both sin
and sin −
.
sin
4
4
4
18. Fill in each blank with an interval of real numbers.
(a) The function f (x) = cos x has domain
and
.
range
and range
(b) The function f (x) = csc x has domain
.
(c) The restricted tangent function has domain
and range
.
(d) The function f (x) = sec−1 x has domain
and
.
range
19. Suppose θ is an angle in standard position whose terminal edge intersects the unit circle at the point (x, y).
1
3
If y = − , what are the possible values of cos θ ? If you
know that the terminal edge of θ is in the third quadrant,
what can you say about cos θ? What if the terminal edge
of θ is in the fourth quadrant? Could the terminal edge of
θ be in the first or second quadrant?
√
function by
20. Show that − 3 is in the range of the tangent
√
finding an angle θ for which tan θ = − 3.
21. Describe restricted domains for sin x, tan x, and sec x on
which each function is invertible. Then describe the corresponding domains and ranges for arcsin x, arctan x, and
arcsec x.
22. Fill in the blanks:
whose
(a) sin−1 x is the angle in the interval
is x.
, for all
(b) y = arcsin x if and only if sin y =
and y ∈
.
x∈
(c) If tan−1 x = θ and tan θ is positive, then θ is in the
quadrant.
1
3
(d) If arctan x = θ and sin θ = , then cos θ =
.
23. Which of the following expressions are defined? Why or
why not?
1
3
(b) sin−1
(a) sin−1 −
25
(c) tan
−1
−1
(d) sec
100
2
π
4
24. Sketch a graph of the restricted cosine function on the
domain [0, π], and argue that this restricted function is
one-to-one. Then sketch a graph of cos−1 x, and list the
domains and ranges of the inverse cos−1 x of this restricted function.
25. Without calculating the exact or approximate values of
the following quantities, use the unit circle to determine
whether each of those quantities is positive or negative:
(a) sin−1 −
1
5
(b) sin−1 −
(c) tan−1 2
2
3
(d) sec−1 (−5)
26. Find all angles whose secant is 2, and then find
sec−1 (2).
Skills
Find the domains of the functions in Exercises 27–32.
ln(x + 1)
27. f (x) =
ln(x − 2)
1
29. f (x) = ln(x − 1)
31. f (x) =
√
sec θ
28. f (x) =
1
e x − e 2x
1
30. f (x) =
1 − tan θ
1
e2
log7 9
+ log3 1
log7 1/3
35. 4 log2 6 − 2 log2 9
36.
π
37. tan −
38. cos
48π
3
40. sin(201π )
42. sin−1 (−1)
41. cos−1 (−1)
2
1
43. arcsec − √
44. arctan − √
3
2
32. f (x) = 2 sin (x − 3)
34. log1/2 4
4
5π
4
−1
Find the exact values of each of the quantities in Exercises
33–44. Do not use a calculator.
33. ln
39. csc −
Solve the equations in Exercises 45–50 by hand. When you are
finished, check your answers either by testing your solutions
or by graphing an appropriate function.
46. 2 = 10 1 +
45. 2 x = 3 x−1
47. log2
x−1
x+1
49. cos 2x = 1
=4
48. sin x =
1
2
50. sec−1 x = π
CONFIRMING PAGES
0.19
12
12x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.4
1
6
3
5
Suppose that cos(θ ) = , sin(θ) > 0, sin(φ) = , and cos(φ) <
0. Use trigonometric identities to identify the quantities in
Exercises 51–56.
51. sin(θ)
use a graphing utility to check that your function f has the
properties and features of the given graph.
y
y
73.
74.
52. sin(−φ)
π
54. sin θ +
53. cos(2θ )
Write each of the expressions in Exercises 57–60 as an
algebraic expression that does not involve trigonometric or
inverse trigonometric functions.
58. tan(tan−1 2x)
59. sec2 (tan−1 x)
60. sin2 (tan−1 x)
61. sin sec−1
3
x
65. f (x) = −
67. f (x) = 20 − 5e
2
3
4
x
4 3 2 1
1
2
2
3
3
4
4
4
2
3
4
1
2
3
x
y
76.
4
2
1
5
1
66. f (x) = −0.25(3
x−2
2
1
x
5 4 3 2 1
1
2
4
y
77.
)
70. f (x) = sin(2x) + 4
72. f (x) = tan−1 (x − 2) + π
x
3
5
y
78.
6
1
5
4
68. f (x) = log1/2 x
69. f (x) = − log2 (x − 3)
π
71. f (x) = 2 cos x −
1
3
x
−2x
1
1
10
3
+ 10
2
15
Sketch graphs of the functions in Exercises 65–72 by hand,
without using a calculator or graphing utility. Indicate any
roots, intercepts, and asymptotes on your graphs.
1
2
3
2
y
75.
62. csc(2 tan−1 x)
x
64. tan2 2 sec−1
63. cos(2 sin−1 5x)
4
3
4 3 2 1
1
56. the sign of tan(θ + π)
57. sin(cos−1 x)
4
1
2
55. the sign of cos(θ + φ)
61
Exponential and Trigonometric Functions
3
π
π
2
π
2
1
π
x
2
1
π
For each graph in Exercises 73–78, find a function whose
graph looks like the one shown. When you are finished,
π
2 1
2
π
2
π
x
Applications
79. Ten years ago, Jenny deposited $10, 000 into an investment account. Her investment account now holds
$22, 609.80. Her accountant tells her that her investment
account balance I(t) is an exponential function.
(a) Find an exponential function of the form I(t) = Ae kt
to model Jenny’s investment account balance.
(b) Find an exponential function of the form I(t) = Ab t
to model Jenny’s investment account balance.
80. Suppose there were 500 rats on a certain island in 1973
and 1,697 rats on the same island 10 years later. Assume
that the number R(t) of rats on the island t years after 1973
is an exponential function.
(a) Find an equation for the exponential function R(t)
that describes the number of rats on the island. Let
t = 0 represent the year 1973.
(b) According to your function R(t), how many rats will
be on the island in 2020?
(c) How long did it take for the population of rats to double from its 1973 amount? How long did it take for it
to double again? And again?
81. Suppose a rock sample initially contains 250 grams of the
radioactive substance unobtainium, and that the amount
of unobtainium after t years is given by an exponential
function of the form S(t) = Ae k t . The half-life of unobtainium is 29 years, which means that it takes 29 years for
the amount of the substance to decrease by half.
(a) Find an equation for the exponential function S(t).
(b) What percentage of unobtainium decays each year?
(c) How long will it be before the rock sample contains
only 6 grams of unobtainium?
82. Again considering the rock sample described in Exercise 81, answer the following questions:
(a) At one point the rock sample contained 900 grams of
unobtainium; how long ago?
(b) What percentage of the unobtainium will be left in
300 years?
(c) How long will it be before 95% of the unobtainium
has decayed?
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
62
Chapter 0
November 26, 2012
Functions and Precalculus
83. Alina is flying a kite and has managed to get her kite so
high in the air that she has let out 400 feet of kite string. If
the angle made by the ground and the line of kite string
is 32 degrees, how high is the kite?
t
ee
0f
40
84. Suppose two stars are each 60 light-years away from
Earth. The angle between the line of sight to the first
star and the line of sight to the second star is 2 degrees.
In other words, if you look at the first star, then turn
your head to look at the second star, your head will move
through an angle of 2 degrees. How far apart are the stars?
s
60 light year
x
2°
x
60 light ye
32°
ars
Proofs
85. Prove by contradiction that every exponential function
f (x) = Ab x has the property that f (x) is never zero. (Hint:
Use what you know about the algebraic properties of exponential functions, and the fact that if f (x) = Ab x is an exponential
function, then neither A nor b is zero.)
86. Use the definition of a one-to-one function to prove
that every exponential function f (x) = Ab x is one-to-one.
(Hint: Use the fact that b x = 1 only when x = 0.)
93. Use the results of the two exericses above to prove that:
1
= − logb x
(a) logb
x
x
= logb x − logb y
(b) logb
87. Use the base conversion formula for logarithms to prove
that the function f (x) = log2 x is equal to the function
g(x) = log3 x only when x = 1.
88. Use logarithms to prove that every exponential function
of the form f (x) = Ab x can be written in the form f (x) =
Ae kx , and vice versa.
95. Use the unit-circle definitions of sine and cosine to prove
the identity sin2 θ + cos2 θ = 1.
96. Use the first Pythagorean identity sin2 θ + cos2 θ = 1 to
prove the second and third Pythagorean identities listed
in Theorem 0.26. (Hint: To prove the second identity, divide
both sides of the first identity by cos2 x. A similar strategy
will prove the third identity.)
89. Use the definition of a logarithmic function y = logb x to
prove that for any b > 0 with b = 1, the quantity logb 1 is
equal to zero.
In Exercises 90–94, assume that x, y, a, and b are values which
make sense in the expressions involved.
90. Use the fact that logarithmic functions are the inverses of
exponential functions to prove that:
(a) logb x = y if and only if b y = x
(b) logb (b x ) = x
(c) blogb x = x
91. Prove that logb (x a ) = a logb x. (Hint: Start with logb (x a )
and replace x with blogb x .)
92. Prove that logb (xy) = logb x + logb y. (Hint: Show that this
statement is equivalent to the statement xy = blog b x+log b y ,
and prove the new statement instead.)
y
94. Prove the base conversion formula logb x =
Set y = logb x and then show that b y = x.)
log a x
. (Hint:
log a b
97. Use the unit-circle definitions of the trigonometric functions to prove the even-odd identities and the shift identities listed in Theorem 0.26.
98. Use the sum identities and the even-odd identities to
prove the difference identities listed in Theorem 0.26.
99. Use the sum identities to prove the double-angle identities listed in Theorem 0.26. (Hint: Note that 2θ is equal
to θ + θ.)
100. The four identities listed as alternative forms in Theorem 0.26 are alternative ways of writing the double-angle
identity cos 2θ = cos2 θ − sin2 θ . Use this double-angle
identity, algebra, and the Pythagorean identities to prove
these four alternative forms.
Thinking Forward
A special exponential limit: Use a calculator to approximate
e h −1
for the following values of h: (a) h = 0.1;
h
(b) h = 0.01; (c) h = 0.001. As h gets closer to
zero, what number does your approximations seem to
approach?
Logarithms with absolute values: Sketch a graph of the
function f (x) = ln |x|. What is the domain of this function? Is the function even, odd, or neither, and why?
Rewriting trigonometric expressions: Use the doubleangle identity sin2 x =
1−cos2x
to rewrite the expres2
sion sin4 x cos2 x in terms of a sum of expressions of the
form A cos kx. (Note: You’ll have to multiply out some
expressions, and use the double-angle identity more
than once.)
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.5
0.5
63
Logic and Mathematical Thinking*
LOGIC AND MATHEMATICAL THINKING*
Logical statements that involve quantifiers or implications
Using a counterexample to show that a statement is false
Mathematical proof techniques, including direct proof and proof by contradiction
From Definitions to Theorems
Throughout this chapter we have learned a lot of the mathematical language that we will
use throughout the book. It is now time to start thinking about how to build on our basic
definitions and develop the theory of calculus. Developing such a theory is like building
a skyscraper. With our library of mathematical definitions and language we have set the
foundation. Throughout the rest of this book we will build on that foundation by using
logical, mathematical arguments to develop new theorems. Each new theorem will be the
springboard to a new set of definitions and theorems and will form a new level of our
skyscraper. In this section we present an overview of the logic and proof techniques needed
for our construction of calculus.
In what follows we will keep the definitions simple so that we can focus on the logic
and proofs. Most of the mathematical statements we will discuss will concern real numbers and whether they are positive, negative, or zero, or will build upon the following five
definitions:
An integer is a whole number that is positive, negative, or zero. The set of all integers
is
{. . . , −5, −4, −3, −2, −1, 0, 1, 2, 3, 4, 5, . . .}.
An integer n is divisible by an integer m if we can write n = km for some integer k.
An integer n is even if we can write n = 2k for some integer k.
An integer n is odd if we can write n = 2k + 1 for some integer k.
A rational number is a real number that can be written in the form
integers p and q, where q = 0.
p
q
for some
Quantifiers
We will often be interested in stating that a property is true all of the time, some of the
time, at least once, or none of the time. Logical quantifiers are the key to making such
statements precisely.
DEFINITION 0.28
The Quantifiers “For All” and “There Exists”
Suppose P is a property that depends on a value x.
(a) For all x, property P means that property P holds for all possible values of x.
(b) There exists x such that property P means that property P is true for at least one
value of x.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
64
Chapter 0
November 26, 2012
Functions and Precalculus
For example, consider the following four quantified statements:
For all real numbers x, 2x + 2 = 2(x + 1).
For all integers x,
1
x
is a rational number.
There exists an integer x such that x is even.
There exists a real number x such that
1
x
= 0.
The first statement is true because its claim that 2x + 2 = 2(x + 1) is indeed true for every
1
real number x. The second statement is false because its claim that is rational fails for
x
x = 0. The third statement is true because many even integers exist; for example, x = 2.
1
The fourth statement is false because there is not even one real number x for which = 0.
x
For statements with more than one quantifier, the order in which the quantifiers appear
can make a big difference. For example, consider the following two statements, which differ
only in the order that their quantifiers are listed:
“For all x > 0, there exists y > 0 such that y > x.”
“There exists y > 0 such that for all x > 0, y > x.”
The first statement claims that given any positive number x, we can find some number y
that is greater than x. This is clearly true, since we can always choose y as large as we like
to ensure that it is larger than the given number x. The second statment claims that there
is some positive number y with the property that y is greater than every number x. This is
clearly false, since no real number is greater than all other real numbers.
Implications
Statements of the form “If A, then B” are called implications. Most of the theorems in this
book have the form of an implication. The statement A is called the hypothesis, and the
statement B is called the conclusion. We will use implications in this book so often that
we have a shorthand notation for them, namely, “A ⇒ B” (pronounced “A implies B”).
DEFINITION 0.29
Implications
An implication is a statement of the form if A, then B (also written A ⇒ B). Such an
implication is true if, whenever statement A is true, statement B must also be true.
For example, the statement “If x > 2, then x > 0” is an implication. In the arrow notation
we would write this as “x > 2 ⇒ x > 0,” and as a quantified statement we could equivalently write “For all x > 2, x > 0.” The hypothesis of the statement is “x > 2,” and the
conclusion is “x > 0.” Thus, if we know that x is greater than 2, then we can conclude that
x must also be greater than 0.
Suppose we have statements A and B, and the statement that A implies B. Does this
mean that B also implies A? Not necessarily; for example, it is true that x > 2 implies x > 0,
but the reverse implication is not true: x > 0 does not imply x > 2. When we switch the
roles of the hypothesis and the conclusion of an implication, we have a new implication
called the converse of the original.
DEFINITION 0.30
The Converse of an Implication
The converse of the implication A ⇒ B is the implication B ⇒ A.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.5
Logic and Mathematical Thinking*
65
If an implication A ⇒ B and its converse B ⇒ A happen to both be true, then we have a
two-way implication. We write this as A ⇐⇒ B, which is pronounced A if and only if B.
For example, the statements “x is even” and “x is divisible by 2” are equivalent; each implies
the other. Thus we can say that “x is even if and only if x is divisible by 2.”
The converse of an implication is obtained when the hypothesis and conclusion switch
places. This results in a very different statement from the original. However, if the hypothesis and conclusion switch places and are negated, then surprisingly, we end up with a
statement that is equivalent to the original.
DEFINITION 0.31
The Contrapositive of an Implication
The contrapositive of the implication A ⇒ B is the statement (Not B) ⇒ (Not A).
The contrapositive of an implication is always logically equivalent to the original implication. For example, the contrapositive of the statement “If x is an integer, then x is a rational
number” is the statement “If x is not a rational number, then x is not an integer.” These
two statements are logically equivalent (and happen to be true).
Counterexamples
A statement of the form “for all x, property P” is false if there is even one instance x for
which P is false. Such an instance is called a counterexample. For example, the statement
“all cats are grey” means that every cat is grey. If there is one cat that is not grey, then some
cats are not grey, and this shows that the statement “all cats are grey” is false.
DEFINITION 0.32
THEOREM 0.33
Counterexamples
A counterexample is an example of a value that makes a statement false.
Counterexamples to “For All” Statements
Suppose P is a property that depends on a value x. Then the statement “For all x, we
have P” is false if and only if there is a counterexample for which P is false.
For example, consider the statement “For all integers x,
1
x
is a rational number.” This state-
ment is false, because we can find a counterexample: when x = 0,
1
x
is not a rational
number. In contrast, the statement “For all real numbers x, 2x + 2 = 2(x + 1)” is true, since
there are no counterexamples x that do not have the property 2x + 2 = 2(x + 1). Finding a
counterexample is a fast and easy way to prove that a “for all” statement is false. To show
that a “for all” statement is true, however, requires substantially more work, as we will see
shortly in this section.
What about counterexamples to implications? An implication “A ⇒ B” implies nothing
about the truth or falsehood of A. Moreover, if A is false, then the implication “A ⇒ B”
does not imply anything about statement B. For example, consider the statement “If I am
elected, then I will lower your taxes.” A politician who says this, gets elected, but then
does not lower taxes is a liar. But that same politician is not a liar if he fails to get elected,
whether he raises taxes or not. A statement of the form “If A, then B” is false only if there
is an instance when its hypothesis A is true and yet its conclusion B is false.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
66
Chapter 0
THEOREM 0.34
November 26, 2012
Functions and Precalculus
Counterexamples to Implications
Suppose A and B are statements that depend on a value x. Then the statement “For all x,
if A, then B” is false if and only if we can find a counterexample for which A is true but
B is false.
For example, the quantified implication “For all x, if x is positive, then x is even” is false,
because we can exhibit a counterexample, say, x = 3, that is positive but not even. Notice
that there are some values of x that are positive and even, but not all values of x that are
positive are also even.
Simple Mathematical Proofs
A mathematical proof is a logical argument. Every theorem in this book can be proved
with the use of previous theorems or definitions. Calculus, like mathematics in general, is
about building up a logical system of definitions, facts, and theorems that can be used to
investigate functions and describe real-world phenomena. In mathematics, the “buildingup” is just as important as any eventual application, and it is very important that each new
theorem rest on a foundation of previous theorems and definitions. It is not enough simply
to rely on our intuition of what ought to be true; we must make sure that every statement we
assert is true and that every theorem that we state is mathematically and logically sound.
If you’re wondering “when you’ll ever use this stuff,” perhaps the best answer is that
you probably won’t, at least not directly. You probably won’t need to directly use your studies of Shakespeare or American history, either. If you pursue a career in science, then you
might use calculus to model or analyze real-world situations. If you become a literary critic,
then you probably won’t need to solve equations, find derivatives, or solve integrals. However, learning calculus and the theory behind it will help teach you how to think. No matter
what you choose to do, the ability to think logically and solve problems will be an invaluable
asset. That is yet another reason that it is so important not only to learn the calculational
mechanics of calculus or how to apply calculus to real-world problems, but also to understand the theory—and thus the proofs—of calculus.
A proof of a statement or theorem of the form “if A, then B” is a logical argument that
starts by assuming the hypothesis A and then argues that the conclusion B must follow.
We assume that A is true, and then make a clear, concise, logical argument that B must
also be true. We indicate that a proof is over by making a box “ ” or by writing “QED,”
which represents the Latin phrase quod erat demonstrandum, meaning “which was to be
demonstrated.”
As a simple example, we present a proof that every integer that is divisible by 10 must be
an even number. Notice that the proof is not much more than a tour through the definitions
of divisibility and even numbers:
Proof. Suppose n is an integer that is divisible by 10. By the definition of divisibility, this means
that we can write n = 10k for some integer k. Rewriting this equation we have n = 2(5k). Since 5k
is also an integer, n satisfies the definition of an even integer. Therefore n is even.
Notice that the proof has a beginning (state the hypothesis), a middle (make a logical
argument), and an end (make the conclusion). This proof is an example of a direct proof,
which means that the conclusion follows from the hypothesis via a fairly straightforward
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.5
Logic and Mathematical Thinking*
67
logical argument. In the examples and exercises we will also explore more involved
methods of proof, such as proof by contradiction, in which we show that something is
true by proving that it cannot be false.
Examples and Explorations
EXAMPLE 1
Determining the truth or falsehood of quantified statements
Determine whether each of the quantified statements that follow are true or false. If a
statement is true, explain why. If a statement is false, provide a counterexample.
(a) For all real numbers x, x ≤ 12.
(b) For all real numbers x, x 2 ≥ 0.
(c) There exists a real number x such that x 2 = −1.
(d) For all integers x, there is some integer y such that y = x + 1.
(e) There exists some integer x such that for all integers y, y = x + 1.
(f) For all e > 0, there exists d > 0 such that for all x > 0, if x < d, then x 2 < e.
SOLUTION
(a) This statement is false, because not all real numbers are less than or equal to 12. The
real number x = 20 is a counterexample.
(b) This statement is true, because the square of any real number is nonnegative. Therefore
there are no counterexamples to this statement.
(c) This statement is false, since
√ no real number has a square that is negative. Although
the complex number i = −1 satisfies x 2 = −1, it is not a real number.
(d) Given any integer x, we can always find some other integer y that is 1 greater than x.
For example, given x = 3, we can choose y = 3 + 1 = 4; given x = 4, we can choose
y = 4 + 1 = 5, and so on. The given statement is true.
(e) There is no integer x for which every other integer y is one greater than x. For example,
for x = 3, there is an integer y that is 1 greater (namely, 3 + 1 = 4), but not all integers
y are one greater than 3. The given statement is false.
(f) This one takes some parsing, but it will be worth it since the given statement is similar
to many of those which we will be studying in Chapter 1. Let’s try an example. If e = 9,
can we find some d such that x < d guarantees that x 2 < 9 for all positive values of x?
2
Yes; if d = 3 and
√x > 0, then x < 3 guarantees that x < 9. In fact, for any value of e,
the value
of d) will make the implication “if
√ d = e (as well as many other values
√
x < e, then x 2 < e” true, since 0 < x < e guarantees that x 2 < e. Therefore the
given statement is true.
EXAMPLE 2
Finding counterexamples to false implications
Each of the implications that follow is false. Provide counterexamples.
(a) If x is even, then x ≥ 0.
(b) The converse of the statement in part (a).
(c) The contrapositive of the statement in part (a).
SOLUTION
(a) This statement is false because not all even integers are nonnnegative. One counterexample is x = −2, since −2 is even but −2 is not greater than or equal to 0.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
68
Chapter 0
November 26, 2012
Functions and Precalculus
(b) The converse of the given statement is “If x ≥ 0, then x is even.” This statement is
false because not all nonnegative integers are even. One counterexample is x = 3,
since 3 ≥ 0 but 3 is not even. In this example the original statement in part (a) and the
converse here in part (b) both happened to be false. In general, however, a statement
and its converse may or may not have the same truth value, since they are logically
different statements.
(c) The contrapositive of the statement in part (a) is “If x < 0, then x is odd.” Notice that
the negation of the statement x ≥ 0 is not the statement x ≤ 0; why? The contrapositive statement is false, because not all negative numbers are odd. One counterexample
is x = −2, since −2 < 0 but −2 is not odd. Notice that we can use the same counterexample for the contrapositive as for the statement in part (a), since the contrapositive
is logically equivalent to the original statement.
EXAMPLE 3
A simple calculational proof
Prove that for all real numbers a and b, a 3 − b 3 = (a − b)(a 2 + ab + b 2 ).
SOLUTION
Sometimes a proof is nothing more than a calculation, written out with justifications for
the steps:
Proof. For any real numbers a and b,
(a − b)(a 2 + ab + b 2 ) = a 3 + a 2 b + ab 2 − a 2 b − ab 2 − b 3
=a −b .
3
← multiply out
← simplify
3
Therefore, a − b = (a − b)(a + ab + b ).
3
EXAMPLE 4
3
2
2
A direct proof
Prove that the sum of any two rational numbers is a rational number.
SOLUTION
Before writing a proof, it is helpful to write down the hypothesis that you are given and
the conclusion that you are trying to show. While doing this, give names to the variables
involved.
Given: a and b are any rational numbers.
Show: a + b is a rational number.
The first line of your proof will be the “given,” the last will be the “show,” and the middle
will be an argument that uses definitions and logical inferences to connect the two. Since
rational numbers are involved, we should remind ourselves of their definition: Recall that
a number is rational if it can be written as the quotient of two integers. This gives us an
updated and more descriptive version of our “given” and “show”:
Given: a =
p
q
and b =
r
s
for some integers p, q, r, and s with q = 0 and s = 0.
Show: a + b can be written in the form
u
v
for some integers u and v with v = 0.
We now have a very clear road map that indicates how we should prove the implication;
we just have to get from the “given” to the “show” by calculating a + b:
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.5
Logic and Mathematical Thinking*
69
Proof. Suppose a and b are rational numbers. Then we can write a and b as quotients of integers,
p
r
say, a = and b = , for some integers p, q, r, and s, where q and s are nonzero. With this notation,
q
s
the sum of a and b is
a+b=
r
ps + qr
p
+ =
.
q
s
qs
Since p, q, r, and s are integers, so are p s + q r and q s. Moreover, since q and s are nonzero, so is
q s. Therefore we have written a + b as a valid quotient of two integers, and thus a + b is a rational
number.
EXAMPLE 5
A proof by contradiction
Prove that the sum of a rational number and an irrational number is irrational.
SOLUTION
Let’s begin by writing out the “given” and “show:”
Given: r is a rational number and x is an irrational number.
Show: r + x is irrational.
A number is irrational if it cannot be written as the quotient of two integers. It can be
difficult to show that we can’t write a number as a quotient of two integers, so instead of
using a direct proof we will use the method of proof by contradiction. This means that
we will show that r + x is irrational by showing that it cannot possibly be rational. More
precisely, we will suppose that r + x is rational and then show that this supposition causes
a logical contradiction.
Proof. Suppose r is a rational number and x is an irrational number. Seeking a contradiction,
suppose that the sum r + x is rational. In the previous example we proved that the sum of two
rational numbers is rational. In addition, if a number r is rational, then so is −r. This means that
the sum of the rational numbers r + x and −r must also be rational, so (r + x) + (−r) = x must be
rational. But this conclusion contradicts our assumption that x is irrational. If r + x is rational, we
are led to a contradiction; therefore r + x must be an irrational number.
TEST YOUR
? UNDERSTANDING
What does it mean to say that a statement A “implies” a statement B? How is it different
than asserting that A and B are true?
What can you say about an implication if its hypothesis is always false? What about if
its conclusion is always true?
Why is x = 3 not a counterexample to the implication “If x is even, then x ≥ 0”?
What does it mean to say that an implication and its contrapositive are logically equiv-
alent statements?
Define each of the following: counterexample, implication, converse, contrapositive.
EXERCISES 0.5
Thinking Back
Basic definitions: Recall the definitions of each of the following
terms or quantities:
nonnegative
integer
rational number
irrational number
|x|
dist(a, b)
Types of numbers: Are all integers rational? Are all
rational numbers integers? Is 0 a rational number?
Why or why not?
Inequality opposites: If x > 9 is false for a particular
value of x, does this necessarily mean that x < 9? Why
or why not? What is the logical opposite of the statement “x > 9”?
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
70
Chapter 0
November 26, 2012
Functions and Precalculus
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: You can show that “For all x, P” is true
by exhibiting just one value of x that makes P true.
(b) True or False: You can show that “For all x, P” is false
by exhibiting just one value of x that makes P false.
(c) True or False: You can show that “There exists x such
that we have P” is true by exhibiting just one value of
x that makes P true.
(d) True or False: You can show that “There exists x such
that we have P” is false by exhibiting just one value
of x that makes P false.
(e) True or False: The converse of an implication is also an
implication.
(f) True or False: When A is true and B is false, the implication A ⇒ B is false.
(g) True or False: When A is false and B is true, the implication A ⇒ B is false.
(h) True or False: When A is false and B is false, the implication A ⇒ B is false.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A true statement involving two quantifiers, and a
false statement involving two quantifiers.
(b) A statement of the form “For all x, A ⇒ B” that has
just one counterexample, and another that has many
counterexamples.
(c) An implication whose converse is false, and an implication whose converse is true.
3. Suppose the implication “C ⇒ D” is true. If C is true,
what can you say about D? If C is false, what can you say
about D?
4. Suppose the implication “R ⇒ S” is false. What does this
mean about statements R and S?
5. Consider the statement “Every positive real number is
greater than −2.” Write this statement using the quantifier “for all.” Then write a statement that is logically
equivalent but uses “if . . . , then . . .” instead of quantifiers.
6. Consider the statement “The square of any real number
is nonnegative.” Write this statement using the quantifier
“for all.” Then write a statement that is logically equivalent but uses “if . . . , then . . .” instead of quantifiers.
7. Consider the statement “Every square is a rectangle.” Is
this statement true? Write down the converse and the
contrapositive of the statement, and determine whether
they are true or false.
8. What is the converse of the statement C ⇒ D? Is the converse logically equivalent to the original statement? Why
or why not?
9. What is the contrapositive of the statement P ⇒ Q? Is the
contrapositive logically equivalent to the original statement? Why or why not?
10. Prove, by exhibiting examples, that the sum of two irrational numbers can be either rational or irrational. Why is
it okay to prove “by example” here, whereas it is not okay
to prove “by example” in general?
For Exercises 11–16, suppose you know the following (no
more and no less) about a function f (x) at values of x:
If 0 < | x − 3| < 0.1, then | f (x) + 5| < 0.2.
Note that this implication means that if we know that x
is a solution of the double inequality 0 < |x − 3| < 0.1, then
we can conclude that f (x) is a solution of the inequality
| f (x) + 5| < 0.2. Use the meaning of the given implication
to determine whether or not each of the following related implications is guaranteed to be true. (Hint: You may have to solve
inequalities or sketch number lines and think about distances to
determine the meanings of the hypotheses and conclusions of the
implications.)
11. If 0 < |x − 3| < 0.1, then | f (x) + 5| < 0.1.
12. If 0 < |x − 3| < 0.1, then | f (x) + 5| < 0.3.
13. If 0 < |x − 3| < 0.05, then | f (x) + 5| < 0.2.
14. If 0 < |x − 3| < 0.05, then | f (x) + 5| < 0.1.
15. If 0 < |x − 3| < 0.05, then | f (x) + 5| < 0.4.
16. If 0 < |x − 3| < 0.2, then | f (x) + 5| < 0.4.
Skills
Determine whether each statement in Exercises 17–46 is true
or false. Justify your answers with reasoning, examples, or
counterexamples, as appropriate.
17. There is some real number between 2 and 3.
18. Every real number is a rational number.
21. For all real numbers x, there is some real number y with
y = x 2.
19. No real number is both rational and irrational.
20. Every real number is either rational or irrational.
24. All real numbers are either greater than zero or less than
zero.
22. For all real numbers x, there is some real number y with
x = y2 .
23. If x is an integer greater than 1, then x ≥ 2.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
0.5
Logic and Mathematical Thinking*
71
25. For all real numbers x, either x is even or x is odd.
26. For all real numbers x, if x < −2, then x 2 > 4.
46. There exists a real number x such that for all real numbers
y, y > x.
27. There exists a real number x such that x ≤ 1 and x ≥ 2.
28. There exist real numbers x and y such that x + y = 4.
Suppose A and B represent logical statements. In Exercises 47–54, write (a) the converse and (b) the contrapositive
of the given statement. Simplify each of your statements if
possible.
29. For all real numbers x, if x is negative, then x is irrational.
30. For all real numbers x, if x is an integer, then x is rational.
31. There exists a real number x such that x ≤ 1 or x ≥ 2.
32. There exists a real number x such that x > 0 and x 2 > 10.
47. (Not A) ⇒ B
48. A ⇒ (Not B)
49. (Not B) ⇒ (Not A)
50. (Not A) ⇒ (Not B)
33. For all real numbers a and b, if a < b, then 3a + 1 < 3b + 1.
51. (A and B) ⇒ C
52. (A or B) ⇒ C
34. No rational number is both less than
53. A ⇒ (B and C)
54. A ⇒ (B or C)
1
3
1
and greater
2
than .
35. For all real numbers x, x 2 ≥ 0 and |x| ≥ 0.
36. For all real numbers x, either x ≥ 2 or x ≤ 1.
37. There exist real numbers x < 0 and y < 0 such that
xy < 0.
38. There exist real numbers x > 0 and y > 0 such that
xy = 0.
39. For all real numbers x and y, if x < y, then 2x − 1 < 2y − 1.
40. For all real numbers x and y, if x < y, then x 2 < y2 .
41. There exists a real number x such that for all real numbers
y, |y| > x.
42. For all real numbers x and y, xy = 0 if and only if x = 0 or
y = 0.
43. For all real numbers x, there exists some y such that
x < y.
44. For all real numbers x, there exists some y such that
x = y2 .
45. For all real numbers x and y,
x
= 0 if and only if x = 0.
y
In Exercises 55–66, write (a) the converse and (b) the contrapositive of each statement. Simplify your statements as much
as possible. (c) Provide counterexamples if the original, the
converse, and/or the contrapositive statements are false.
55. If x is a real number, then x is rational.
56. If x ≥ 2, then x ≥ 3.
57. If x > 2, then x ≥ 3.
58. If x ≥ 2, then x ≥ 1.
√
59. If x is negative, then x is not a real number.
60. If x is rational, then x is not irrational.
61. If x ≤ 0, then |x| = −x.
62. If x < −2, then |x| = −x.
63. If x is not zero, then x 2 > x.
64. If x is positive and rational, then x − 1 is positive and
rational.
65. If x is odd, then there is some integer n such that
x = 2n + 1.
66. If x is even, then there is some integer n such that x = 2n.
Applications
Use logic to solve the puzzles in Exercises 67 and 68. Then
write proofs to argue that your solutions are correct.
67. Linda, Alina, Phil, and Stuart are wearing differentcolored hats: either red, yellow, green, or blue. From the
following statements, determine which hat each person is
wearing:
Neither boy wears a red hat.
The oldest person wears a green hat.
Linda is older than Alina.
Alina never wears yellow or red.
Stuart is the youngest and hates blue.
68. Xena, Yolanda, and Zeke each have different favorite
fruits: either apples, bananas, or cantaloupes. Use the
statements that follow to determine which person prefers
which fruit. Then write a proof that argues that your
solution is correct.
Xena likes bananas better than apples.
Zeke is allergic to cantaloupes.
Bananas are the favorite fruit of one of the girls.
Yolanda likes bananas better than cantaloupes.
Consider an Island X where there are exactly two types
of people: those who always tell the truth (“truth-tellers”)
and those who always lie (“liars”). Given the statements in
Exercises 69–71, determine which people are truth-tellers and
which are liars. Then write a proof which argues that your
solution is correct. (These types of puzzles, based on those made
popular by Raymond Smullyan, are a great avenue to developing
your logical thinking skills and your proof-writing abilities.)
69. You meet Liz, Rein, and Zubin, who say the following:
Liz: “We all tell the truth.”
Rein: “Exactly two of us tell the truth.”
Zubin: “Liz and Rein always lie.”
70. You meet Anita, Bill, and Chris, who say the following:
Anita: “At least one of Bill or Chris tells the truth.”
Bill: “Anita is a liar.”
Chris: “Bill is a liar.”
71. You meet Hyun, Jaan, and Kate, and only Kate and Hyun
have something to say:
Kate: “Hyun and Jaan are liars.”
Hyun: “Kate always tells the truth.”
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
72
Chapter 0
November 26, 2012
Functions and Precalculus
Proofs
72. Prove that x 2 − y2 = (x − y)(x + y) for all real numbers
x and y.
73. Prove that if x is an irrational number and r is a rational
number, then the difference x − r must be an irrational
number. You may assume that the sum of two rational
numbers is a rational number.
74. Prove that for any real numbers a and b, |b − a| = |a − b|.
75. Prove that the sum of an even integer and an odd integer
is odd.
76. Prove that the sum of an odd integer and an odd integer
is even.
77. Prove that the product of an even integer and an odd
integer is even.
78. The Pythagorean theorem states that if a right triangle has
legs of length a and b, and a hypotenuse of length c, then
a 2 + b 2 = c2 . Use the Pythagorean theorem and the definition of the distance between two real numbers to prove
that the distance between any two points P = (x1 , y1 ) and
Q = (x2 , y2 ) in the plane is given by the “distance for
mula” (x2 − x1 )2 + ( y2 − y1 )2 . (Hint: Draw an example of
two points P and Q in the plane, label their coordinates, and
use an appropriate right triangle.)
x +x y +y
79. Prove that the midpoint 1 2 , 1 2 between the
2
2
points P = (x1 , y1 ) and Q = (x2 , y2 ) is equidistant from
P and Q.
80. Prove that the numbers
x=
−b +
√
b 2 − 4ac
2a
and x =
−b −
√
b 2 − 4ac
2a
found by the quadratic formula are solutions of the
quadratic equation ax 2 + bx + c = 0.
81. Use the fact that
(a/b)
ad
(a/b)
a
=
to prove that
= . You
(c/d)
bc
c
bc
may assume that all denominators are nonzero.
a
c
ac
a
ac
=
= .
to prove that c
82. Use the fact that
b
d
bd
b
b
83. Follow the steps outlined here to prove the triangle inequality: |a + b| ≤ |a| + |b| for any real numbers a
and b:
√
(a) Argue that for any real number x, |x| = x 2 .
(b) Show that (a + b)2 ≤ (|a| + |b|)2 . (Hint: Start on the
left-hand side, multiply out the expression, and use the
fact that a ≤ |a| and b ≤ |b|.)
(c) Take the square root of both sides of the inequality
from part (b) (this is valid since both sides are positive), and use part (a) to show that |a + b| ≤ |a| + |b|.
Use the triangle inequality to prove the following two inequalities, for any real numbers a and b:
84. |a − b| ≤ |a| + |b|
85. |a − b| ≥ |a| − |b|
For Exercises 86 and 87, use the definition of absolute value
and systems of inequalities to prove that for any real numbers
x and c, and any positive real number δ, the given statement
is true:
86. |x − c| < δ ⇐⇒ x ∈ (c − δ, c + δ)
87. |x − c| > δ ⇐⇒ x ∈ (−∞, c − δ) ∪ (c + δ, ∞)
88. The “Monty Hall Problem” is a good example of a problem about which people’s initial intuition is often incorrect. On the game show Let’s Make a Deal, the host, Monty
Hall, presents you with a choice of three doors. Behind
one door is a lot of money. Behind the other two doors
are worthless gag prizes. You pick a door and point at it.
Monty Hall knows which door conceals the prize, and he
opens one of the two doors you didn’t pick to show you a
gag prize. Then he gives you the option of keeping your
original choice or switching your choice to the remaining
door.
(a) Is it better to switch or to stick with your original
choice? Or are both choices equally likely to lead to
the money? Think about this problem for awhile and
convince yourself of an answer before you go on to
the next part.
(b) Come up with a proof or argument that would convince another person of the correct answer.
You may assume that any denominators are nonzero.
Thinking Forward
Quantified statements about distance: Show that each of the
statements that follow is true by exhibiting a value of δ that
satisfies each statement. These types of statements will be the
backbone of our study of limits in Chapter 1. (Note: The symbols and δ are Greek letters that represent real numbers that
are usually positive and quite small. In the third statement you
will need to write δ in terms of .)
There exists a δ > 0 such that for all x, if x is within
distance δ of 2, then 3x + 1 is within distance 1 of 7.
There exists a δ > 0 such that for all x, if |x − 2| < δ,
then |3x − 6| < 0.3.
For all > 0, there exists a δ > 0 such that if x is within
distance δ of 3, then 2x is within distance of 6.
Negating quantified statements about distance: Write down the
negation of each statement that follows. We will see these
types of statements again when we formally define the concept of a limit in Chapter 1.
For all M > 0, there exists an N > 0 such that for all x,
if x > N, then x 2 > M.
For all > 0, there exists a δ > 0 such that for all x, if
0 < |x − 2| < δ, then |x 2 − 4| < .
For all > 0, there exists
√ a δ > 0 such that for all x, if
0 < |x − 4| < δ, then | x − 2| < .
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
Chapter Review, Self-Test, and Capstones
73
CHAPTER REVIEW, SELF-TEST, AND CAPSTONES
Before you progress to the next chapter, be sure you are familiar with the definitions, concepts, and basic skills outlined here.
The capstone exercises at the end bring together ideas from this chapter and look forward to future chapters.
Definitions
Give precise mathematical definitions or descriptions of each
of the concepts that follow. Then illustrate the definition with
a graph or algebraic example, if possible.
the form of a rational function
what it means for a function to be algebraic
the piecewise-defined form of f (x) = |x|, and more generally of a function f (x) = |g(x)|
what it means for a function to be transcendental
the form of an exponential function, and when this type of
function represents exponential growth or exponential decay
the natural exponential function and the natural logarithmic
function
the algebraic definitions of even functions and odd functions, and the graphical meaning of y-axis symmetry and
rotational symmetry
what it means for functions f and g to be inverses of each
other
a function from a set A to a set B
the independent variable and the dependent variable of a
function
the domain and range of a function, in set notation
the graph of a function, in set notation and as a picture
a one-to-one function
a root of a function and a y-intercept of a function
a local maximum, local minimum, global maximum, or global
minimum of a function
an inflection point of a function (in rough terms)
what it means for a function f to be positive, negative,
increasing, or decreasing on an interval I
the sine and cosine of an radian angle θ , in terms of coordinates on the unit circle.
what it means (roughly) for a function f to be concave up
or concave down on an interval I
the average rate of change of a function f on an interval
[a, b]
the tangent, cotangent, secant, and cosecant of a radian
angle θ, in terms of the sine and cosine of θ
a piecewise-defined function
the inverse sine, inverse tangent, and inverse secant functions and their domains and ranges
the meaning of the quantified statement “For all x, P”
the form of a power function and the form of a polynomial
function
the meaning of the quantified statement “There exists x
such that P”
the leading coefficient, the leading term, and the constant
term of a polynomial
the meaning of the implication “If A, then B”
the hypothesis and the conclusion of an implication
the forms of constant, linear, quadratic, cubic, quartic, and
quintic polynomials
the converse and the contrapositive of an implication A ⇒ B
a counterexample to a statement
Fill in the blanks to complete each of the following statements
of theorems:
If f is an invertible function with inverse f −1 , then the
and the range of f −1 is
.
domain of f −1 is
y = f (x) + C is shifted C units
from y = f (x) if C > 0
from y = f (x) if C < 0.
and C units
y = f (x + C) is shifted C units
from y = f (x) if C > 0
from y = f (x) if C < 0.
and C units
If f is an invertible function with inverse f −1 , then
f −1 (b) = a if and only if
, and the graph of y = f −1 (x)
.
can be obtained from the graph of y = f (x) by
A function has an inverse if and only if the function is
.
If f is a polynomial function of degree n, then the graph
real roots and at most
local
of f has at most
extrema.
If f is a polynomial function, then the graph of f behaves
at its “ends.”
like the graph of
what it means for a quadratic polynomial to be irreducible,
and how this is related to the discriminant
Theorems
y = k f (x) is vertically stretched from y = f (x) by a factor
and vertically compressed by a factor of k if
of k if
.
y = f (k x) is horizontally stretched from y = f (x) by a facand horizontally compressed by a factor of
tor of k if
.
k if
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
74
Chapter 0
If f (x) =
November 26, 2012
Functions and Precalculus
p(x)
is a rational function, then f is not defined
q(x)
If f (x) =
p(x)
is a rational function, then f has holes at the
q(x)
, provided that
points that are roots of
.
p(x)
If f (x) =
is a rational function, then f has vertical
q(x)
, provided
asymptotes at the points that are roots of
.
that
f (x) =
p(x)
q(x)
is a rational function with
deg( p(x)) = n and deg(q(x)) = m. If n < m, then f
; if n = m, then f has
has a horizontal asymptote at
; and if n > m, then f
a horizontal asymptote at
.
, and f has roots at the points that are
at the roots of
but not roots of
.
roots of
Suppose
A statement of the form “For all x, P” is false if and only
.
if there is a counterexample in which
A statement of the form ”For all x, if A, then B” is false if
but
and only if there is a counterexample in which
.
Notation and Algebraic Rules
Notation: Describe the meanings of each of the following
mathematical expressions.
f:A→B
y = f (x)
f (x) = |x|
(k f )(x)
( f + g)(x)
( f · g)(x)
( f ◦ g)(x)
A ⇐⇒ B
f
g
(x)
A⇒B
Trigonometric identities: Fill in the blanks to complete each of
the following trigonometric identities, where θ , α, and β are
angles measured in radians.
sin2 θ + cos2 θ =
tan2 θ + 1 =
( f ◦g ◦h)(x)
1 + cot 2 θ =
sin(−θ ) =
cos(−θ ) =
tan(−θ ) =
π
=
2
π
=
2
Logarithms: Fill in the blanks to complete each of the algebraic
rules that follow. You may assume that x, y, b, and a are real
numbers whose values make the expressions well-defined.
cos θ −
logb x = y ⇐⇒
logb (b x ) =
sin(θ + 2π) =
cos(θ + 2π) =
blogb x =
logb (x a ) =
sin(α + β) =
cos(α + β) =
logb (xy) =
logb
sin(α − β) =
cos(α − β) =
sin 2θ =
cos 2θ =
logb
x
y
1
x
=
loga x
=
logb x
=
sin θ +
Skill Certification: Algebra and Functions
Simplifying expressions: Simplify each expression as much as
possible.
1
1
−
2+h
2
x 3 − 23
1.
x−2
2.
x−2
3.
x 3 − x 2 − 4x + 4
x + 27x
4.
x 2 + 5x + 6
5. |−2(x 2 + 1)|
6.
7. f (x) =
h
15.
4
3
<1
x−2
17. |5x − 2| > 1
|4 − 2x|
x−2
8. f (x) =
9. f (x) = e2 ln x
π π + tan
11. f (x) = tan
3
13. 2x 2 − 7x + 3 > 0
4
x1/4 + x1/3
x2
Solving inequalities: Solve each of the inequalities, and express
each solution set in interval notation.
x2 − 9
≤0
x−1
1
16. |3x − 4| <
2
14.
18. |x 2 − 4| < 2
Finding zeros and undefined values: Determine the x-values for
which each function is zero, and the x-values for which each
function does not exist.
√
x −2/5 4x
√
3
x
10. f (x) = log2 (8(4 x ))
1 2
12. f (x) = sin−1 −
2
19. f (x) =
2x 2 − 5x + 3
x
21. f (x) = |x − 2| − 5
20. f (x) = 3x 4 − 6x 3 + x 2
22. f (x) =
CONFIRMING PAGES
1
1
−
x
1−x
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 26, 2012
75
Chapter Review, Self-Test, and Capstones
23. f (x) =
2(x − 1)2 − 4
x−1
2x 3 − 1
2
2x − 5x − 3
x ln x
26. f (x) = x
e −1
24. f (x) =
25. f (x) = e (1 − 2e )
x
x
1
28. f (x) =
sec x + 1
sin(π x)
27. f (x) =
x
1
29. f (x) =
sin−1 x
43. f (x) =
√
x−3
45. f (x) = 2x 3 − 1
47. f (x) =
1
−2
x
44. f (x) = (x − 3)2 + 1
46. f (x) =
48. f (x) =
√
1−x
1
x−2
49. f (x) = |4x − 3|
50. f (x) = |x 2 − 9|
51. f (x) = −2x1/3
52. f (x) = 2x −1/3
Finding domains: Find the domain of each function, and express the domain in interval notation.
√
√
x+2
31. f (x) =
32. f (x) = x 2 − 2x − 3
x
53. f (x) = −x(x + 1)2
54. f (x) = 3x 3 + x 2 − 3x − 1
x+2
33. f (x) = 2
x −x−6
57. f (x) = 2 x
58. f (x) =
59. f (x) = e3x
60. f (x) = e−3x
61. f (x) = 1 − 5e x
62. f (x) = −3e x−1
63. f (x) = ln x
64. f (x) = log2 x
65. f (x) = log1/2 x
66. f (x) = 1 − 3 ln x
67. f (x) = sin x
68. f (x) = tan x
π
70. f (x) = 2 sin x −
30. f (x) = arctan(x 2 + 1)
1
34. f (x) = √
x 2 − 3x
x−3
36. f (x) =
1−x
x + 1
38. f (x) = ln
√
√
35. f (x) = x + 2 − x
37. f (x) =
39. f (x) =
1
e−1/2x
x−1
tan−1 x
40. f (x) =
π − sec−1 x
1
1
− sin x
2
Graphs of basic functions: Sketch the graph of each function by
hand, using your knowledge of simple graphs and transformations. Label any important points or features.
41. f (x) = 3 − 2x
55. f (x) =
(x − 1)(x + 2)
(x − 1)2
69. f (x) = sec x
−1
71. f (x) = sin
56. f (x) =
x3 + x2 + x + 1
2x 2 + x − 1
x
1
2
4
x
−1
72. f (x) = tan
x
42. f (x) = 4(x − 1) + 2
Capstone Problems
A.
Transformations of cubic functions: Prove algebraically that
a vertical or horizontal shift or stretch of a cubic function is also a cubic function. That is, prove that if f (x) is a
cubic function, then so are f (x) + C, f (x + C), k f (x), and
f (k x).
B.
Peeking forward to derivatives: Suppose f (x) = , and
x
consider the two-variable function
1
q(x, h) =
f (x + h) − f (x)
.
h
(a) Simplify q(3, h) as much as possible, and argue that
1
it approaches − as h gets closer to 0.
9
(b) Simplify q(x, h) as much as possible, and argue that
it approaches −
1
as h gets closer to 0.
x2
C.
Approximating the area of a region in the plane: Sketch the
region that lies between the graph of f (x) = 20 − 2 x , the
x-axis, and the lines y = 2 and y = 4. Use geometric figures to approximate upper and lower bounds for the area
of this region.
D. Optimizing the area of a region given its perimeter: Elizabeth wants to build a rectangular pen for her dogs with
100 feet of spare fencing.
(a) Write down an equation in terms of length l and
width w for the perimeter P of the pen.
(b) Use the perimeter equation and the constraint on
fencing material to construct a one-variable equation
for the area A of the enclosure.
(c) Sketch a graph of the one-variable function A, and
use the graph to argue that its maximum occurs
when l = w = 25 and the enclosure is square.
CONFIRMING PAGES
18:20
TKmaster2010
WHF00153/FREE087-Taalman
November 27, 2012
this page left intentionally blank
22:3
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
C H A P T E R 1
Limits
1.1
An Intuitive Introduction to Limits
Examples of Limits
Limits of Functions
Infinite Limits, Limits at Infinity, and Asymptotes
Examples and Explorations
1.2
Formal Definition of Limit
Formalizing the Intuitive Definition of Limit
Uniqueness of Limits
One-Sided Limits
Infinite Limits and Limits at Infinity
Examples and Explorations
1.3
Delta-Epsilon Proofs*
Describing Limits with Absolute Value Inequalities
Finding a Delta for Every Epsilon
Writing Delta-Epsilon Proofs
Examples and Explorations
1.4
∀ > 0, ∃δ > 0 . . .
Continuity and Its Consequences
Defining Continuity with Limits
Types of Discontinuities
Continuity of Very Basic Functions
Extreme and Intermediate Values of Continuous Functions
Examples and Explorations
1.5
Limit Rules and Calculating Basic Limits
Limits of Combinations of Functions
Limits of Algebraic Functions
Finding Limits by Cancelling or Squeezing
Defining the Number e
Continuity of Exponential and Trigonometric Functions
Delta-Epsilon Proofs of the Limit Rules
Examples and Explorations
1.6
lim( f (x) + g(x))
x→c
Infinite Limits and Indeterminate Forms
Infinite Limits
Limits at Infinity
Indeterminate and Non-Indeterminate Forms
Special Trigonometric Limits
Examples and Explorations
Chapter Review, Self-Test, and Capstones
CONFIRMING PAGES
∞
∞
15:42
TKmaster2010
WHF00153/FREE087-Taalman
78
1.1
Chapter 1
November 21, 2012
Limits
AN INTUITIVE INTRODUCTION TO LIMITS
Examples of the types of limits that are important in calculus
Intuitive descriptions of two-sided and one-sided limits
Infinite limits, limits at infinity, and horizontal and vertical asymptotes
Examples of Limits
Limits are the backbone of calculus. Limits are the key to investigating the local behavior of
functions, giving meaning to the slope of a curve, finding areas under curves and volumes
inside surfaces, and determining the long-term behavior of infinite sequences and sums. In
the next section we will formally and mathematically define limits; in this section we focus
on intuitive examples. Let’s start with three examples that illustrate how limits can arise.
Limits of sequences: As a starting point for thinking about the concept of a limit, consider
the following sequence of numbers:
1 1 1 1
1
1
1
, , ,
,
,
, ..., k, ...
2 4 8 16 32 64
2
If the pattern of this sequence continues, then the numbers will continue
to
get smaller and
1
smaller, approaching zero. We say that 0 is the limit of the sequence k as k approaches
2
infinity. We can never actually let k be equal to infinity, because infinity is not a real number.
1
However, we can let k get as large as we like. For each large value of k, the value of k is very
2
small, but not actually zero. When we “take the limit,” we make an important theoretical
1
transition: Instead of evaluating k at a particular value of k, we think about the behavior
2
of the sequence as k gets larger and larger. We think about what the sequence approaches,
even if it never actually gets there for any real number k.
Limits of sequences of sums: Now consider the sequence defined by adding up more and
more terms from the previous sequence:
1
,
2
1 1
+ ,
2 4
1 1 1
+ + ,
2 4 8
1 1 1
1
+ + + ,
2 4 8 16
1 1 1
1
1
+ + + + , ...
2 4 8 16 32
After computing the sums, this sequence is equal to
1 3 7 15 31 63
, , ,
,
,
, ...
2 4 8 16 32 64
The terms get closer and closer to 1 as we go further and further out in this sequence of
sums. You may have noticed a pattern in the sequence: It turns out that for any given k,
the value of
1
2
+
1
4
+ ··· +
1
2k
will be equal to the quantity
2k − 1
.
2k
The larger the value of k,
the closer this quantity gets to 1. Moreover, we can get the sum to be as close as we like
to 1 by choosing a sufficiently large value of k. Again, we can’t plug in infinity for k, since
we can’t in real life add up infinitely many numbers, even if those numbers are getting
infinitesimally small, as they are here. However, mathematically we can use the concept of
a limit to notice that this sequence of sums approaches the quantity 1. We will study limits
of sequences of sums in depth in Chapter 7; for now we present this type of limit only as
an interesting example to consider.
Limits of average rates of change: Now let’s switch gears and think of something a little more
practical. Suppose you drop a bowling ball from the top of a 100-foot parking deck and
want to know how fast it is falling when it hits the ground. Suppose that the height of
the ball t seconds after it is dropped is given by s(t) = 100 − 16t 2 feet and thus that
the ball hits the ground after 2.5 seconds. By the old formula “distance equals rate times
s
time,” the average rate that the ball falls over any time period t will be r = , where s
t
is the elapsed distance over the period t. A good approximation of the final speed of the
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
79
An Intuitive Introduction to Limits
bowling ball is the average rate in the last half-second, which is
r=
s(2.5) − s(2)
0 − 36
=
= −72 feet per second
2.5 − 2
0.5
(negative since the ball is falling downwards). To get a better approximation for the ball’s
final instantaneous velocity, we could calculate the rate over the last quarter-second. The
following table records a sequence of better and better approximations for the final velocity
of the bowling ball:
Time interval
[2, 2.5]
[2.25, 2.5]
[2.4, 2.5]
[2.49, 2.5]
Average rate
ft
−72
sec
ft
−76
sec
ft
−78.4
sec
ft
−79.84
sec
[2.4999, 2.5]
−79.9984
ft
sec
s
We can’t compute the actual final velocity of the ball this way, because we can’t use r =
t
when t is zero. But the average rates seem to approach −80 feet per second as t gets
smaller and smaller. This sounds like another limit, and it is. In fact it is a very famous and
useful type of limit called a derivative that we will introduce in Chapter 2. In general, limits
help us discuss what happens when we let things get infinitesimally small, infinitely large,
or arbitrarily close to some number.
Limits of Functions
Intuitively, a limit is what the output of a function approaches as we let the input of that
function approach some value. In the previous examples, we saw that:
1
approaches 0.
2k
1
1
1
k approaches infinity, the sum + + · · · + k approaches 1.
2
4
2
s(2.5) − s(t)
t approaches 2.5, the average rate
approaches −80 feet
2.5 − t
As k approaches infinity, the quantity
As
As
per second.
If the values of a function f (x) approach some number L as x gets closer and closer to some
value x = c, we will write
lim f (x) = L.
x→c
We can also consider limits of functions as x → ∞, that is, as x grows without bound. For
example, in this notation we have
lim
1
k→∞ 2 k
=0
and
lim
t→2.5
s(2.5) − s(t)
= −80.
2.5 − t
When considering a limit, it only matters what happens as x gets closer and closer to c,
not what happens when it actually gets there. This means that lim f (x) may or may not in
x→c
general be the same as the value f (c) of the function at x = c. For example, the functions
f (x) = x + 1 and g(x) =
x2 − 1
x−1
shown in the following figures are not equal at the point
x = 1, but they do approach the same value as x → 1:
lim f (x) = 2 and f (1) = 2
lim g(x) = 2 but g(1) is undefined
x→1
x→1
y
1
y
3
3
2
2
1
1
1
2
x
1
1
CONFIRMING PAGES
2
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
80
Chapter 1
November 21, 2012
Limits
The dots on the x-axis represent a sequence of values of x that approach x = 1. When
we evaluate the functions f and g at these values, in both cases we get a sequence
of values of y that gets closer and closer to y = 2. Although f (1) = g(1), we do have
lim f (x) = lim g(x) = 2.
x→1
x→1
We can also consider limits as x grows without bound, and/or as f (x) grows without bound.
The following definition summarizes the notation we will use in each case:
DEFINITION 1.1
Intuitive Description of and Notation for Limits
Suppose f is a function and L and c are real numbers.
(a) Limit: If the values of a function f (x) approach L as x approaches c, then we say
that L is the limit of f (x) as x approaches c and we write
lim f (x) = L.
x→c
(b) Limit at Infinity: If the values of a function f (x) approach L as x grows without
bound, then we say that L is the limit of f (x) as x approaches ∞ and we write
lim f (x) = L.
x→∞
(c) Infinite Limit: If the values of a function f (x) grow without bound as x approaches c,
then we say that ∞ is the limit of f (x) as x approaches c and we write
lim f (x) = ∞.
x→c
(d) Infinite Limit at Infinity: If the values of a function f (x) grow without bound as
x grows without bound, then we say that ∞ is the limit of f (x) as x approaches ∞
and we write
lim f (x) = ∞.
x→∞
When a limit approaches a real number, we say that the limit exists. When a limit
approaches ∞ or −∞ we say that the limit does not exist (because ∞ and −∞ are not
real numbers), but we will always be as specific as possible and describe the sign of infinity
in such cases. Later we will see more pathalogical limits that “do not exist” in a way other
than being infinite.
When we say that x approaches a real number c, we implicitly mean to consider values
of x that are close to c from either the right or the left. In other words, when trying to
find lim f (x), we consider both values of x that are slightly less than c as well as values of
x→c
x that are slightly greater than c. Sometimes it is convenient to consider these two cases
separately:
DEFINITION 1.2
Intuitive Description of One-Sided Limits
If the values of a function f (x) approach a value L as x approaches c from the left, we say
that L is the left-hand limit of f (x) as x approaches c and we write
lim f (x) = L.
x→c−
If the values of a function f (x) approach a value R as x approaches c from the right, we
say that R is the right-hand limit of f (x) as x approaches c and we write
lim f (x) = R.
x→c+
Note that the notation x → c− does not mean anything about whether c is a positive or
negative number, only that x approaches c from the left. The two-sided limit of f (x) as x → c
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
81
An Intuitive Introduction to Limits
exists if and only if the left and right limits as x approaches c exist and are equal. This means
that both the left and right limits approach the same real number.
For example, the function graphed here has a different limit from the left than from the
right as x approaches 1:
lim f (x) = 2 but lim f (x) = 3
x→1−
x→1+
y
3
2
1
1
1
x
2
The purple sequence of values of x that approach x = 1 from the left determines a sequence
of values of f (x) that approach y = 2, while the red sequence determines values that approach y = 3. The value of the function at x = 1 happens to be f (1) = 3, but that is
not relevant to either limit calculation. Since the limits from the left and right are not the
same, there is no one real number that the function approaches as x → c and we say the
two-sided limit does not exist.
Infinite Limits, Limits at Infinity, and Asymptotes
Armed with the concept of limits, we can now give proper definitions for horizontal and
vertical asymptotes. If a function f increases or decreases without bound as x approaches a
real number c from either the right or the left, then f has a vertical asymptote at x = c:
DEFINITION 1.3
Vertical Asymptotes
A function f has a vertical asymptote at x = c if one or more of the following are true:
lim f (x) = ∞,
x→c+
lim f (x) = ∞,
lim f (x) = −∞,
x→c−
x→c+
or
lim f (x) = −∞.
x→c−
If f (x) approaches ∞ from both the left and the right as x → c, then we say that lim f (x) =
x→c
∞, as happens in the leftmost graph that follows. If f (x) approaches −∞ from both the left
and the right,then we say that lim f (x) = −∞. If f (x) approaches different signs of infinity
x→c
from the left and the right, then the two-sided limit lim f (x) does not exist, as happens in
x→c
the middle graph.
lim f (x) = ∞
lim f (x) = −∞, lim+ f (x) = ∞
x→1−
x→1
y
3
2
2
1
1
2
y
y
3
2 1
1
lim f (x) = 10
x→∞
x→1
1
2
3
4
x
2 1
1
2
15
10
1
2
3
4
5
x
1
1
5
CONFIRMING PAGES
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
82
Chapter 1
November 21, 2012
Limits
If the values of a function f (x) approach a real-number value as x increases or decreases
without bound, then f has a horizontal asymptote. For example, the rightmost graph of the
three just presented shows a function with a horizontal asymptote and its corresponding
limit. In general, we have the following definition:
DEFINITION 1.4
Horizontal Asymptotes
A nonconstant function f has a horizontal asymptote at y = L if one or both of the
following are true:
lim f (x) = L, or
lim f (x) = L.
x→∞
x→−∞
Note that by convention, if f (x) is actually equal to L as x → ∞ or as x → −∞, then we do
not consider f to have a horizontal asymptote at y = L. For example, the constant function
f (x) = 2 has lim f (x) = 2 and yet does not have a horizontal asymptote at y = 2, since f (x)
x→∞
is constantly equal to 2 as x → ∞.
Examples and Explorations
EXAMPLE 1
Determining limits with tables of values
x
Use tables of values to find (a) lim (x + 1) and (b) lim
x→∞ x − 1
x→1
.
SOLUTION
(a) To see what happens to x + 1 as x → 1, we choose a sequence of values approaching
x = 1 from the left and a sequence approaching x = 1 from the right, and record the
corresponding values of x + 1:
x
.9
.99
.999
1
1.001
1.01
1.1
x+1
1.9
1.99
1.999
*
2.001
2.01
2.1
From both the left and the right, the values of of x + 1 approach 2. Assuming that this
pattern continues for values of x that are even closer to 1, we have lim (x + 1) = 2.
x→1
(b) To see what happens to
x
x−1
as x → ∞, we choose a sequence of values of x that gets
larger and larger, and record the corresponding (rounded) values of
x
x
x−1
25
50
100
1000
10,000
1.04167
1.02041
1.0101
1.001
1.0001
As x grows larger, the quantity
x
x−1
approaches 1, so, assuming that the pattern in the
x
x→∞ x − 1
table continues for even larger values of x, we have lim
EXAMPLE 2
x
:
x−1
= 1.
Graphically identifying limits
Determine the limits at any holes, corners, or asymptotes on the graphs of the functions
(a) f , (b) g, and (c) h:
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
y = f (x)
y = g(x)
y
3
2
2
1
1
3 2 1
1
y = h(x)
y
y
3
1
2
3
x
3 2 1
1
2
2
3
3
83
An Intuitive Introduction to Limits
3
2
1
2
3
x
1
3 2 1
1
2
3
x
SOLUTION
(a) Observe that the graph in question has holes at x = −2 and x = −1 and a corner at
x = 1, so we will examine limits at those three points. As x approaches −2, the height
of the graph of y = f (x) approaches −1. The value of f (x) at x = −2 is f (−2) = 1,
but this is not relevant to what f (x) approaches as x → −2 and thus does not affect the
limit. Therefore lim f (x) = −1.
x→−2
As x approaches −1, the height of the graph approaches −2. The value f (−1) is not
defined, but that is not relevant to the limit as x → −1. Thus lim f (x) = −2.
x→−1
As x approaches 1, the height of the graph approaches 2. The value of f (x) at x = 1
happens to also equal 2, although this is irrelevant to the value of the limit as x → 1.
We have lim f (x) = 2.
x→1
(b) The function g(x) approaches different values as we approach x = 1 from the left and
the right. As x approaches 1 from the left, the height of the graph approaches y = 2. As
x approaches 1 from the right, the height of the graph approaches y = 1. The value of
g(x) at x = 1 happens to be g(1) = 1, but that is not relevant to either limit. Therefore
we have lim− g(x) = 2 and lim+ g(x) = 1, but the two-sided limit lim g(x) does not
exist.
x→1
x→1
x→1
(c) The function h(x) has a vertical asymptote at x = −1, with the height of the function
decreasing without bound as we approach from the left and increasing without bound
as we approach from the right. Therefore we have lim − h(x) = −∞ and lim + h(x) =
x→−1
∞, but the limit lim h(x) does not exist.
x→−1
x→−1
Let’s also investigate the limits at the ends of the graph of h(x). On the left side, as
x → −∞, the graph decreases without bound; therefore lim h(x) = −∞. On the
x→−∞
right side, as x → ∞, the height of the function approaches y = 1. Therefore h(x) has
a horizontal asymptote on the right, and lim h(x) = 1.
x→∞
EXAMPLE 3
A function with infinitely many oscillations as x approaches 0
Use a table of values and various graphing windows
on
a calculator or other graphing utility
to investigate the limit of the function f (x) = sin
1
x
as x approaches 0.
SOLUTION
Be sure that your calculator or other graphing utility is set to radian mode,
rather than degree mode, for this example. To see what happens to the quantity sin
1
x
CONFIRMING PAGES
as x approaches 0,
15:42
TKmaster2010
WHF00153/FREE087-Taalman
84
Chapter 1
November 21, 2012
Limits
we choose points progressively closer to x = 0 from both the left and the right and record
1
x
the corresponding (rounded) values of sin
x
1
sin
x
in a table:
−0.001
−0.0001
−0.00000001
0
0.00000001
0.0001
0.001
−0.827
0.306
−0.932
??
0.932
−0.306
0.827
From the table we see that as x approaches 0, the
of f (x) seem to jump around! It is
values
not clear whether or not the function f (x) = sin
1
x
will eventually approach any particular
1
from this table.
value as x → 0. It is impossible to make an educated guess for lim sin
x
x→0
1
oscillates
The reason that this is happening is that as x → 0 the function f (x) = sin
x
faster and faster between −1 and 1, and never settles down. The graph on the left that
follows shows this function on [−3, 3], and the graph on the right shows the same function
after reducing by a factor of 10 on the x-scale (but keeping the y-scale the same). No matter
how much
we “zoom in” towards x = 0, this function will keep oscillating. Therefore
lim sin
x→0
1
x
does not exist.
f (x) = sin
1
x
f (x) = sin
on [−3, 3]
1
x
y
y
1
3
2
on [−0.3, 0.3]
1
1
1
2
3
x
0.3
0.1
0.2
1
0.1
0.2
0.3
x
1
EXAMPLE 4
Areas under curves as another application of limits
Consider the area between the graph of f (x) = x 2 and the x-axis from x = 0 to x = 1, as
shown in the leftmost figure.
(a) Find the value of the four-rectangle approximation shown in the middle figure.
(b) Find the value of the eight-rectangle approximation shown in the rightmost figure.
(c) Describe what would happen if we were to do similar approximations with more and
more rectangles.
Area under f (x) = x 2 on [0, 1]
Four rectangles
y
Eight rectangles
y
y
1.00
1.00
1.00
0.75
0.75
0.75
0.50
0.50
0.50
0.25
0.25
0.25
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
0.25
CONFIRMING PAGES
0.50
0.75
1.00
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
85
An Intuitive Introduction to Limits
SOLUTION
1
(a) The four rectangles in the given middle graph each have width , and their heights are
4
1
2
3
4
,f
,f
, and f
, from left to right. Therefore, since f (x) = x 2 ,
given by f
4
4
4
4
the sum of the areas of the four rectangles is
2 2 2 2 1
1
2
1
3
1
4
1
15
+
+
+
=
≈ 0.4688.
4
4
4
4
4
4
4
4
32
1
(b) Similarly, the eight rectangles in the given rightmost graph each have width , and
8
1
2
3
8
their heights are given by f
,f
,f
, ...,f
, from left to right. Therefore
8
8
8
8
the sum of their eight areas is
2 2 2 2 1
1
2
1
3
1
8
1
51
+
+
+ ··· +
=
≈ 0.3984.
8
8
8
8
8
8
8
8
128
(c) So what happens when we do similar approximations with more rectangles? Consider
the following three figures, where we consider more and more rectangles:
16 rectangles
32 rectangles
y
64 rectangles
y
y
1.00
1.00
1.00
0.75
0.75
0.75
0.50
0.50
0.50
0.25
0.25
0.25
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
Clearly as we let the number N of rectangles get larger and larger, the sum of the areas
of the rectangles gets closer and closer to a particular real number, namely, the actual
area under the graph of f (x) = x 2 on [0, 1]. In limit notation we have
lim (N-rectangle area approximation) = (actual area under graph).
N→∞
It turns out that the area approximations corresponding to the 16-, 32-, and
64-rectangle figures are 0.3652, 0.3491, and 0.3412, respectively. In Chapter 7 we will
develop theory that will allow us to show that these area approximations have a limit
1
of as N → ∞.
3
TEST YOUR
? UNDERSTANDING
What is the difference between taking a limit of some quantity as, say, x approaches 2
and actually computing the value of the quantity at x = 2?
What can you say about the two-sided limit of a function as x → c if the left and right
limits as x → c both exist but are not equal to each other?
Can a function have more than one horizontal asymptote? More than two? Use
Definition 1.4 to support your answer.
Explain why the table in Example 1 cannot guarantee that lim (x + 1) is actually 2 rather
x→1
than some other number that is close to 2, such as 2.000035.
1
does not exist?
Why does the table in Example 3 suggest that lim sin
x→0
x
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
86
Chapter 1
November 21, 2012
Limits
EXERCISES 1.1
Thinking Back
Finding the pattern in a sequence: For each sequence shown, find
the next two terms. Then write a general form for the kth term
of the sequence.
2, 6, 10, 14, 18, 22, . . .
3,
3 3 3 3 3
, , , , ,...
4 9 16 25 36
3 4 5 6 7 8
, , , , , ,...
5 7 9 11 13 15
1, 8, 27, 64, 125, 216, . . .
1,
1 1 1 1 1
, , , ,
,...
3 9 27 81 243
3 5 7 9 11 13
, , , , , ,...
2 5 10 17 26 37
Distance, rate, and time: A watermelon dropped from the top
of a 50-foot building has height given by s(t) = 50 − 16t 2 feet
after t seconds. Calculate each of the following:
The average rate of change of the watermelon over
its entire fall, over the first half of its fall, and over the
second half of its fall.
The average rate of change over the last second, the
last half-second, and the last quarter-second of its fall.
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: A limit exists if there is some real number that it is equal to.
(b) True or False: The limit of f (x) as x → c is the value
f (c).
(c) True or False: The limit of f (x) as x → c might exist
even if the value f (c) does not.
(d) True or False: The two-sided limit of f (x) as x → c
exists if and only if the left and right limits of f (x)
exist as x → c.
(e) True or False: If the graph of f has a vertical asymptote
at x = 5, then lim f (x) = ∞.
x→5
(f) True or False: If lim f (x) = ∞, then the graph of f has
x→5
a vertical asymptote at x = 5.
(g) True or False: If lim f (x) = ∞, then the graph of f has
x→2
a horizontal asymptote at x = 2.
(h) True or False: If lim f (x) = 2, then the graph of f has
4. If lim+ f (x) = −2, lim− f (x) = 3, and f (0) = −2, what can
x→0
x→0
you say about lim f (x)?
x→0
5. If lim+ f (x) = 8 but lim f (x) does not exist, what can you
x→2
x→2
say about lim− f (x)?
x→2
6. If lim + f (x) = −∞ and lim − f (x) = −∞, what can you
x→−1
x→−1
say about lim f (x)?
x→−1
7. If lim f (x) = ∞, lim f (x) = 3, and lim+ f (x) = ∞, what
x→−∞
x→∞
1 2 3 4
2 3 4 5
8. Consider the sequence , , , , . . .,
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function f for which f (2) does not exist
but lim f (x) does exist.
x→2
(b) The graph of a function f for which f (2) exists and
lim f (x) exists, but the two are not equal.
x→2
(c) The graph of a function f for which neither f (2) nor
lim f (x) exist.
x→2
3. If lim− f (x) = 5 and lim+ f (x) = 5, what can you say about
x→1
x→1
lim f (x)? What can you say about f (1)?
x→1
k
, . . ..
k+1
(a) What happens to the terms of this sequence as k gets
larger and larger? Express your answer in limit notation.
(b) Use a calculator to find a sufficiently large value of k
so that every term past the kth term of this sequence
will be within 0.01 unit of 1.
1 1 1 1
1
, , . . ., k , . . ..
3 9 27 81
3
9. Consider the sequence , ,
x→−∞
a horizontal asymptote at y = 2.
x→1
can you say about any horizontal and vertical asymptotes
of f ?
(a) What happens to the terms of this sequence as k gets
larger and larger? Express your answer in limit notation.
(b) Find a sufficiently large value of k so that every term
past the kth term of this sequence will be less than
0.0001.
10. Consider the sequence of sums
1
1
1
1
+ +
+ , . . ..
3
9
27
81
1 1
1 1
1
1
, + , + + ,
3 3
9 3
9
27
(a) What happens to the terms of this sequence of sums
as k gets larger and larger?
(b) Find a sufficiently large value of k which will guarantee that every term past the kth term of this sequence
of sums is in the interval (0.49999, 0.5).
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
11. Consider the sequence of sums 1, 1 + 2, 1 + 2 + 3,
1 + 2 + 3 + 4, 1 + 2 + 3 + 4 + 5, . . ..
15. Consider the area between the graph of f (x) = 4 − x 2 and
the x-axis on [0, 2].
(a) What happens to the terms of this sequence of sums
as k gets larger and larger?
(b) Find a sufficiently large value of k that will guarantee
that every term past the kth term of this sequence of
sums is greater than 1000.
12. An orange falling from 20 feet has a height of s(t) =
20 − 16t 2 feet when it has fallen for t seconds.
(a) Graph the position function s(t) and find the time that
the orange will hit the ground.
(b) Make a table to record the average rates that the orange is falling during the last second, half-second,
quarter-second, and eighth-of-a-second of its fall.
(c) From the data in your table, make a guess for the
instantaneous final velocity of the orange at the
moment it hits the ground.
13. If you are on the moon, then an orange falling from
20 feet has a height of s(t) = 20 − 2.65t 2 feet when it
has fallen for t seconds.
(a) Graph the position function s(t) and find the time that
the orange will hit the surface of the moon.
(b) Make a table to record the average rates that the orange is falling during the last second, half-second,
quarter-second, and eighth-of-a-second of its fall on
the moon.
(c) From the data in your table, make a guess for the
instantaneous final velocity of the orange at the
moment it hits the surface of the moon.
√
14. Consider the area between the graph of f (x) = x and
the x-axis on [0, 4].
y
y
2
y
y
4
4
3
3
2
2
1
1
x
2
1
2
1
x
(a) Use the four rectangles shown on the left to approximate the given area, and then use the eight rectangles
shown on the right to obtain another approximation
of that area. Be sure to use the fact that the graph
shown is that of the function f (x) = 4 − x 2 in your
calculations.
(b) Describe what would happen if we did similar approximations with more and more rectangles, and
make a guess for the resulting limit.
16. Sketch a function that has the following table of values,
but whose limit as x → ∞ is equal to −∞:
x
100
200
500
1,000
10,000
f (x)
50
55
56.2
56.89
56.99
17. Sketch a function that has the following table of values,
but whose limit as x → 2 does not exist:
x
1.9
1.99 1.999 2 2.001 2.01
f (x) 3.12 3.09
3.01
-
2.99
2.1
2.92 2.87
18. Use a calculator or other graphing utility to graph the
function f (x) =
2
1
87
An Intuitive Introduction to Limits
x−2
.
x2 − x − 2
(a) Show that f (x) is not defined at x = 2. How is this
reflected in your calculator graph?
(b) Use the graph to argue that even though f (2) is
1
1
3
undefined, we have lim f (x) = .
x→2
1
2
3
4
x
1
2
3
4
x
(a) Use the four rectangles shown on the left to approximate the given area, and then use the eight rectangles
shown on the right to obtain another approximation
of that area. Be sure to use the fact √
that the graph
shown is that of the function f (x) = x in your calculations.
(b) Describe what would happen if we did similar approximations with more and more rectangles, and
make a guess for the resulting limit.
19. Use a calculator or other graphing utility to graph the
function g(x) =
x 2 − 2x + 1
.
x−1
(a) Show that g(x) is not defined at x = 1. How is this
reflected in your calculator graph?
(b) Use the graph to argue that even though g(1) is
undefined, we have lim g(x) = 0.
x→1
20. Use a calculator or other
graphing utility to investigate the
graph f (x) = x sin
1
x
near x = 0. Be sure to have your
calculator set to radian mode. Use the graphs to make an
educated guess for lim f (x).
x→0
Skills
Sketch the graphs of functions that have the given limits
and values in Exercises 21–30. (There are multiple correct
answers.)
21.
lim f (x) = 3 and lim f (x) = −∞
x→−∞
x→∞
22. lim f (x) = −4 and lim f (x) = −∞
x→2
x→−∞
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
88
23.
24.
25.
26.
27.
28.
29.
30.
Chapter 1
November 21, 2012
Limits
lim f (x) = ∞ and lim− f (x) = ∞
x→0+
Use tables of values to make educated guesses for each of the
limits in Exercises 39–52.
x→0
lim f (x) = 3 and lim+ f (x) = 1
x→5−
x→5
39.
lim f (x) = 2 and lim f (x) = 2
lim f (x) = 2, lim+ f (x) = 2, but f (3) does not exist
1
x2 − 4
x−3
43. lim 2
x→3 (x − 2)(x − 3)
lim f (x) = −2 and lim f (x) = ∞, f (0) = −5
45. lim
lim f (x) = 2, lim+ f (x) = −1, and f (2) = 2
47.
lim f (x) = 3, lim+ f (x) = 3, and f (2) = 0
49.
x→−∞
x→∞
lim f (x) = ∞, lim+ f (x) = −∞, and f (2) = 1
x→2−
x→2
x→3−
x→3
x→−∞
x→2
x→2
x→2−
x→2
51.
y
3
2
1
4 3 2 1
1
1
2
3
4
x
2
3
32.
33.
34.
lim f (x), lim + f (x), lim f (x), and f (−2).
x→−2−
x→−2
x→−2
lim − f (x), lim + f (x), lim f (x), and f (−1).
x→−1
x→−1
x→−1
lim− f (x), lim+ f (x), lim f (x), and f (2).
x→2
x→2
x→1
lim
lim
x→∞
x→−∞
x→∞
y
3
2
1
1
2
3
4
x
2
x+1
x2 − 1
36.
37.
38.
x→−1
x→−1
lim g(x), lim+ g(x), lim g(x), and g(1).
x→1−
x→1
x→1
lim g(x), lim+ g(x), lim g(x), and g(2).
x→2−
x→2
x→2
lim g(x), lim g(x), lim g(x), and lim g(x).
x→0
x→3
x→−∞
1 + 2x
x−1
1
lim 1 +
lim
x→∞
x→∞
2x + 1
+
1
x2
lim sin x
52.
x→∞
lim sin
x→∞
1
x
Sketch graphs by hand and use them to make approximations
for each of the limits in Exercises 53–66. If a two-sided limit
does not exist, describe the one-sided limits.
1
54. lim (x 3 − 2)
53. lim
x→−1
x→0 x
x2 − 1
x2 + x − 2
56. lim
55. lim
x→−2
x→1 x − 1
x+2
x−1
x−4
58. lim 2
57. lim 2
x→∞ x − 4
x→1 x − 1
59.
61.
lim (1 − e −x )
60.
lim tan x
62. lim csc x
x→∞
63. lim f (x), for f (x) =
x→2
64. lim f (x), for f (x) =
x→0
lim (3e 4x + 1)
x→−∞
x→π
x→π/2
x 2 , if x < 2
1 − 3x, if x ≥ 2
2x + 1, if x ≤ 0
2x − 1, if x > 0
⎧
⎨ x 2 + 1,
3,
65. lim f (x), for f (x) =
⎩
x→1
3 − x,
⎧
⎨ x + 1,
2,
66. lim f (x), for f (x) =
⎩
x→−1
−x 2 ,
67. lim(3 − 4x − 5x 2 )
lim − g(x), lim + g(x), lim g(x), and g(−1).
x→−1
50.
lim (3e −2x + 1)
x→∞
if x < 1
if x = 1
if x > 1
if x < −1
if x = −1
if x > −1
Use calculator graphs to make approximations for each of the
limits in Exercises 67–74.
3
35.
48.
For the function g(x) graphed as follows, approximate each of
the limits and values in Exercises 35–38:
4 3 2 1
1
46.
3x + 1
1−x
x→2
lim f (x), lim f (x), lim f (x), and lim f (x).
x→0
1
1−x
x−5
44. lim 2
x→5 x − 25
x→1
3
4 − 2x
x→−∞
lim (1 − 3x + x 2 )
x→3+
42. lim
x→2
x→3
x→2−
40.
41. lim
For the function f graphed as follows, approximate each of
the limits and values in Exercises 31–34:
31.
lim (x 2 + x + 1)
x→2−
x→∞
x→4
3−x
x−1
x100
71. lim x
x→∞ 2
sin x
73. lim
x→0 x
69. lim
x→1
68.
lim (−0.2x 5 + 100x)
x→∞
x+1
x−2
ln x
72. lim
x→∞ x
1 − cos x
74. lim
x→0
x
70. lim
x→2
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.1
89
An Intuitive Introduction to Limits
Applications
75. There are four squirrels currently living in Linda’s attic. If
she does nothing to evict these squirrels, the number of
squirrels in her attic after t days will be given by the formula S(t) =
12 + 5.5t
.
3 + 0.25t
(a) Use the graph to approximate the temperature of the
yam when it is first put in the oven.
(b) Use the graph to approximate lim T(t).
t→∞
(c) What is the temperature of the oven, and why?
(a) Verify that there are four squirrels in Linda’s attic at
time t = 0.
(b) Determine the number of squirrels in Linda’s attic after 30 days, 60 days, and one year.
(c) Approximate lim S(t) with a table of values. What
77. In 1960, H. von Foerster suggested that the human
population could be measured by the function
P(t) =
179 × 109
.
(2027 − t)0.99
t→∞
does this limit mean in real-world terms?
(d) Graph S(t) with a graphing utility, and use the graph
to verify your answer to part (c).
76. The following graph describes the temperature T(t) of a
yam in an oven, where temperature T is measured in degrees Fahrenheit and time t is measured in minutes:
Temperature of yam
T
400
350
300
250
200
150
100
50
Here P is the size of the human population. The time t is
measured in years, where t = 1 corresponds to the year
1 a.d., time t = 1973 corresponds to the year 1973 a.d.,
and so on.
(a) Use a graphing utility to graph this function. You will
have to be very careful when choosing a graphing
window!
(b) Use the graph you found in part (a) to approximate
lim − P(t).
t→2027
(c) This population model is sometimes called the
doomsday model. Why do you think this is? What
year is doomsday, and why?
(d) In part (b), we considered only the left limit of P(t) as
x → 2027. Why? What is the real-world meaning of
the part of the graph that is to the right of t = 2027?
10 20 30 40 50 60
t
Proofs
1
78. Prove that for all k > 100, the quantity 2 is in the
k
interval (0, 0.0001). What
does this have to do with the
limit of the sequence
1
k2
as k → ∞?
x→1
79. For any positive integer k, the following equation holds:
1 + 2 + 3 + ··· + k =
80. Prove that for all x within 0.01 of the value x = 1, the
quantity (x − 1)2 is within the interval (0, 0.0001). What
does this have to do with lim (x − 1)2 ?
k(k + 1)
. Use this fact to prove that
2
for all k > 100, the value of the sum of the first k integers
is greater than 5000. What does this have to do with the
limit of a sequence of sums as k → ∞?
81. Prove that for all x within 0.01 of the value x = 1, the
1
quantity
is greater than 10, 000. What does this
2
(x − 1)
have to do with lim
x→1
1
?
(x − 1)2
Thinking Forward
Convergence and divergence of sequences: If a sequence
a 1 , a 2 , a 3 , . . . , a k , . . . approaches a real-number limit as
k → ∞, then we say that the sequence {a k } converges. If the
terms of the sequence do not get arbitrarily close to some real
number, then we say that the sequence diverges. Write out
enough terms of each sequence to make an educated guess
as to whether it converges or diverges.
k
k
1
4
5
4
k
k+2
k+1
k
Convergence and divergence of series: A series can be thought
of as an infinite sum a 1 + a 2 + a 3 + a 4 + · · · + a k + · · · . A
series converges if this sum gets closer and closer to some
real number limit as we add up more and more terms.
Otherwise, the series is said to diverge.
As you will see in Chapter 8, the series 1 +
1
1
+ +
4
9
1
1
+ · · · + 2 + · · · converges. Calculate partial sums
16
k
including more and more terms until you are convinced that the sum eventually approaches a realnumber limit and does not grow without bound.
Although you might think that the series 1 +
1
+
2
1
1
1
+ + · · · + + · · · converges because its terms
3
4
k
get smaller and smaller, you will see in Chapter 8
that it does not. Calculate partial sums including more
and more terms until you are convinced that this sum
diverges and in fact grows without bound, never
approaching a real-number limit. (A calculator will
come in handy here!)
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
90
1.2
Chapter 1
November 21, 2012
Limits
FORMAL DEFINITION OF LIMIT
Moving from an intuitive concept of limit to a formal mathematical definition
Uniqueness and existence of limits
Limits from the left and right, limits at infinity, and infinite limits
Formalizing the Intuitive Definition of Limit
In the previous section we gave an intuitive description of limits. Now that we understand
the basic concept, we are ready to give a precise, rigorous mathematical definition. Let’s
start with our intuitive description: For real numbers c and L and a function f , we have
lim f (x) = L if the values of f (x) get closer and closer to L as x gets closer and closer to c.
x→c
For example, lim x 2 = 4 because the values of f (x) = x 2 approach 4 as x approaches 2. From
x→2
the left, f (1.9) = (1.9)2 = 3.61, f (1.99) = (1.99)2 = 3.9601, f (1.999) = (1.999)2 ≈ 3.996,
and so on, getting closer and closer to 4. A similar thing happens as x approaches 2 from
the right.
Note that to be able to discuss lim f (x), we must know how to calculate f (x) near, but
x→c
not necessarily at, the point x = c. Throughout this section we will assume that f (x) is
defined on a punctured interval (c − δ, c) ∪ (c, c + δ), where δ > 0 represents a small
distance to the left and right of x = c, as shown on the number line that follows. Notice
that in our discussion of limits we will never be concerned with what happens at the point
x = c, only near the point x = c.
Punctured δ-interval around c
δ
cδ
δ
c
cδ
To make the definition of limit precise, we have to be very clear about what we mean
when we say that f (x) “approaches” L. We want to capture the idea that we can make
the values of f (x) not just close to L, but as close as we like to L if only we choose values of x that are sufficiently close to c. For example, we can guarantee that f (x) = x 2 is
within 0.05 unit of 4 if we choose values of x that are within 0.01 unit of 2. Note that
f (2.01) = (2.01)2 = 4.0401 and f (1.99) = (1.99)2 = 3.9601 are both within 0.05 unit of 4,
and values of x that are closer to 2 will result in values of f (x) that are even closer to 4. If we
want values of f (x) that are even closer to L = 4, then we can just choose values of x that
are even closer to c = 2.
In general, suppose we want to guarantee that the values of f (x) are within some very
small distance above or below limit value L, as shown at the left in the graphs that follow.
To do this we must choose values of x that are sufficiently close to c, say, some distance
δ > 0 left or right of c, as shown in the middle graph. The Greek letters delta (δ) and
epsilon () are the traditional letters used for these small distances. The figure at the right
illustrates that a choice of x-value inside the blue punctured δ-interval (c − δ, c) ∪ (c, c + δ)
determines an f (x)-value within the beige -interval (L − , L + ). In these figures we have
omitted the point at x = c to emphasize that we are not concerned with the actual value of
f (x) at the point x = c, only with the behavior of f (x) at points near x = c.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.2
If x ∈ (c − δ, c) ∪ (c, c + δ),
then f (x) ∈ (L − , L + )
So choose x
in (c − δ, c) ∪ (c, c + δ)
Want to have f (x)
in (L − , L + )
y
y
L
L
L
L
L
L
x
c
91
Formal Definition of Limit
y
L
f(x)
x
cδ c cδ
xc
x
If the values of f (x) get arbitrarily close to L as x approaches c, then we can choose smaller
and smaller beige -intervals (L − , L + ) and in each case always find some blue punctured δ-interval (c − δ, c) ∪ (c, c + δ) that determines values of f (x) which are within of the
limit L. The following three figures illustrate this idea:
y
y
L
y
L
c
x
L
c
x
c
x
We want the limit statement lim f (x) = L to mean that no matter how small a distance
x→c
we choose for , we can find some δ so that values of x that are within δ of x = c will yield
values of f (x) that are within of y = L. Writing this in terms of intervals gives us the
following definition:
DEFINITION 1.5
Formal Definition of Limit
The limit lim f (x) = L means that for all > 0, there exists δ > 0 such that
x→c
if x ∈ (c − δ, c) ∪ (c, c + δ), then f (x) ∈ (L − , L + ).
Stop and think about that for a minute until it makes sense. Understanding this definition
is the key to understanding limits, and limits are the foundation of everything in calculus.
So take a few minutes, have some tea, and get everything straight in your head before you
continue reading.
Uniqueness of Limits
A limit lim f (x) exists if it is equal to some real number L. If a limit exists, then it can be
x→c
equal to one and only one number. That sounds obvious, since the values of a function f (x)
cannot approach two different values L and M as x approaches c. However, as we are
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
92
Chapter 1
November 21, 2012
Limits
about to see, to prove uniqueness of limits we must carefully apply the formal definition of
limits.
THEOREM 1.6
Uniqueness of Limits
If lim f (x) = L and lim f (x) = M, then L = M.
x→c
x→c
Proof. Suppose to the contrary that somehow lim f (x) = L and lim f (x) = M, with L = M. Let’s
x→c
x→c
suppose that L > M, since, if not, then we can just reverse the roles of L and M. If L > M, then we
k
must have L = M + k for some positive real number k. Now consider = ; note that with this
2
choice of , the intervals (L − , L + ) and (M − , M + ) do not overlap.
Since lim f (x) = L, we can find δ 1 > 0 such that for all x ∈ (c − δ 1 , c) ∪ (c, c + δ 1 ), we have
x→c
f (x) ∈ (L − , L + ). Similarly, since lim f (x) = M, we can find δ 2 > 0 such that for all x ∈
x→c
(c − δ 2 , c) ∪ (c, c + δ 2 ), we have f (x) ∈ (M − , M + ). Now if we let δ be the smaller of δ 1 and
δ 2 , we can say that for any x ∈ (c − δ, c) ∪ (c, c + δ), we can guarantee that both f (x) ∈ (L − , L + )
and f (x) ∈ (M − , M + ). But this cannot be, since the intervals (L − , L + ) and (M − , M + ) do
not overlap. Therefore we could not have initially had f (x) approaching two different limits L and
M; we must have L = M.
One-Sided Limits
We can consider each limit lim f (x) = L from two different directions: from the left and
x→c
from the right. We say that we have a left limit lim− f (x) = L if, given an -interval (L − ,
x→c
L + ), we can always find a sufficiently small half-neighborhood (c − δ, c) to the left of
x = c so that values of x that are in that left hand δ-interval yield values of f (x) that are in
the -interval, as shown in the left-hand graph that follows. We define right limits similarly,
as shown in the right-hand graph:
Can get f (x) in (L − , L + )
by choosing x in (c − δ, c)
Can get f (x) in (L − , L + )
by choosing x in (c, c + δ)
y
y
L
L
L
L
L
L
cδ c
DEFINITION 1.7
x
c cδ
x
One-Sided Limits
The left limit lim− f (x) = L means that for all > 0, there exists δ > 0 such that
x→c
if x ∈ (c − δ, c), then f (x) ∈ (L − , L + ).
The right limit lim+ f (x) = L means that for all > 0, there exists δ > 0 such that
x→c
if x ∈ (c, c + δ), then f (x) ∈ (L − , L + ).
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.2
93
Formal Definition of Limit
A two-sided limit lim f (x) is equal to some real number L if and only if the correspondx→c
ing left and right limits exist and are also equal to that same number L.
THEOREM 1.8
For a Limit to Exist, the Left and Right Limits Must Exist and Be Equal
lim f (x) = L if and only if lim− f (x) = L and lim+ f (x) = L.
x→c
x→c
x→c
The proof of this theorem is a straightforward application of the definitions of two-sided
and one-sided limits.
Proof. Suppose lim f (x) = L. Then for all > 0, there is some δ > 0 such that if x ∈ (c − δ, c) ∪
x→c
(c, c + δ), then f (x) ∈ (L − , L + ). In particular this means that if x ∈ (c − δ, c) or if x ∈ (c, c + δ),
then we will have f (x) ∈ (L − , L + ). Therefore lim− f (x) = L and lim+ f (x) = L.
x→c
x→c
For the converse, suppose lim− f (x) = L and lim+ f (x) = L. Then for all > 0, there exist numx→c
x→c
bers δ 1 > 0 and δ 2 > 0 such that for either x ∈ (c−δ 1 , c) or x ∈ (c, c+δ 2 ), we have f (x) ∈ (L−, L+).
If we let δ be the smaller of δ 1 and δ 2 , then we can say that for x ∈ (c − δ, c) ∪ (c, c + δ), we can
guarantee that f (x) ∈ (L − , L + ). Therefore lim f (x) = L.
x→c
Infinite Limits and Limits at Infinity
So far we have formalized the definition of limit only in the case where both x and f (x)
are approaching real numbers. Now we consider what happens if one or both of x and f (x)
approach ±∞. For example, we want lim f (x) = ∞ to capture the idea that as x approaches
x→c
c, the values of f (x) grow without bound. In other words, lim f (x) = ∞ should guarantee
x→c
that values of f (x) will lie above any given large number M as long as we choose values of
x that are sufficiently close to c; see the figure that follows at the left.
lim f (x) = ∞
lim f (x) = L
x→c
lim f (x) = ∞
x→∞
y
x→∞
y
y
M
M
L
L
L
cδ
c
cδ
x
x
N
N
x
Similarly, we want the limit statement lim f (x) = L to indicate that given any
x→∞
-interval around L, we can choose values of x sufficiently large so that f (x) is in the
-interval; see the middle graph. And lim f (x) = ∞ should mean that given any large
x→∞
number M, we can get values of f (x) that are greater than M if we choose sufficiently large
values of x, say, larger than some big number N, as in the graph at the right. The definition
that follows expresses these three limits in terms of intervals; compare the three parts with
the preceding three figures. You will define other related types of limits in Exercises 37–42.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
94
Chapter 1
DEFINITION 1.9
November 21, 2012
Limits
Limits Involving Infinity
The infinite limit lim f (x) = ∞ means that for all M > 0, there exists δ > 0 such that
x→c
if x ∈ (c − δ, c) ∪ (c, c + δ), then f (x) ∈ (M, ∞).
The limit at infinity lim f (x) = L means that for all > 0, there exists N > 0 such
x→∞
that
if x ∈ (N, ∞), then f (x) ∈ (L − , L + ).
The infinite limit at infinity lim f (x) = ∞ means that for all M > 0, there exists
x→∞
N > 0 such that
if x ∈ (N, ∞), then f (x) ∈ (M, ∞).
Examples and Explorations
EXAMPLE 1
Approximating δ given for a limit
Use a graph to illustrate and approximate
(a) the largest δ which guarantees that if x ∈ (2 − δ, 2) ∪ (2, 2 + δ), then x 2 ∈ (3, 5).
(b) the largest δ which guarantees that if x ∈ (2 − δ, 2) ∪ (2, 2 + δ), then x 2 ∈ (3.5, 4.5).
What limit statement do these problems have to do with, and why?
SOLUTION
These problems concern the limit statement lim x 2 = 4, which by definition means that for
x→2
all > 0, there exists δ > 0 such that if x ∈ (2 − δ, 2) ∪ (2, 2 + δ), then x 2 ∈ (4 − , 4 + ).
Therefore parts (a) and (b) of this example ask us to find corresponding values of δ for = 1
and = 0.5, respectively.
(a) To find the largest δ corresponding to = 1, we begin by drawing f (x) = x 2 and the
beige -interval of width 1 around y = 4, as shown in the graph that follows at the
left. We then draw the vertical blue band shown in the figure, to represent the range of
values of x which determine values of f (x) that are within the horizontal beige band.
The leftmost x-value a in the blue band is a solution of f (a) = a2 = 3, and the
√ rightmost
x-value b√for the blue band is a solution of f (b) = b2 = 5. Therefore a = 3 ≈ 1.732
and b = 5 ≈ 2.236. Now, what is δ in this case? We can move 2 − 1.732 = 0.268 unit
to the left of x = 2 and 2.236−2 = 0.236 unit to the right of x = 2. We need the largest δ
that will work in both directions, which is the smaller of the two distances we just found:
δ = 0.236. Then if x ∈ (2 − 0.236, 2) ∪ (2, 2 + 0.236), we can guarantee that x 2 ∈ (3, 5).
If x ∈ (1.732, 2.236), then x 2 ∈ (3, 5)
If x ∈ (1.87, 2.12), then x 2 ∈ (3.5, 4.5)
y
y
5
4.5
4
3.5
4
3
1.732
2
2.236
x
1.87
2
CONFIRMING PAGES
2.12
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.2
95
Formal Definition of Limit
(b) We now repeat the problem with a smaller value of . For = 0.5 we draw the smaller
horizontal beige -bar shown in the preceding graph at the right, and the corresponding vertical blue bar. Solving f (a) = a2 = 3.5 and f (b) = b2 = 4.5, we get a ≈ 1.87
and b ≈ 2.12 as the leftmost and rightmost values of x contained in the vertical blue
bar. Therefore values of x that are at most 2 − 1.87 = 0.13 unit to the left of x = 2 or
at most 2.12 − 2 = 0.12 unit to the right of x = 1 will determine values of f (x) that are
within 0.5 unit of y = 4. The smaller of these two distances is the largest δ that will
work in both directions, namely, δ = 0.12. If x ∈ (2 − 0.12, 2) ∪ (2, 2 + 0.12), then we
can guarantee that x 2 ∈ (3.5, 4.5).
EXAMPLE 2
Approximating N given for a limit at infinity
Use a graph to illustrate and approximate
(a) the smallest N which guarantees that if x ∈ (N, ∞), then
(b) the smallest N which guarantees that if x ∈ (N, ∞), then
x+1
x
x+1
x
∈ (0.75, 1.25).
∈ (0.9, 1.1).
What limit statement do these problems have to do with, and why?
SOLUTION
These problems are about the limit statement lim
x→∞
x+1
x
= 1, which by definition means
that for all > 0, there is some N > 0 such that if x ∈ (N, ∞), then
x+1
x
∈ (1 − , 1 + ).
Therefore parts (a) and (b) of this example ask us to find the corresponding values of N for
= 0.25 and = 0.1, respectively.
x+1
(a) The figure that follows at the left shows f (x) =
and a beige bar representing all the
x
heights within 0.25 unit of y = 1. The blue area shows all of the values of x for which
the corresponding values of f (x) lie within the beige bar. According to this graph, to
find the leftmost point x = a of the blue area we must solve f (a) = 1.25:
a+1
= 1.25 =⇒ a + 1 = 1.25a =⇒ 1 = 0.25a =⇒ a = 4.
a
f (a) = 1.25 =⇒
Therefore if x ∈ (4, ∞), then we can guarantee that
If x ∈ (4, ∞), then
x+1
∈ (0.75, 1.25)
x
x+1
x
∈ (0.75, 1.25).
If x ∈ (10, ∞), then
y
x+1
∈ (0.9, 1.1)
x
y
1.25
1.1
1
1
0.75
0.9
4
x
10
x
(b) We now do the same thing but for = 0.1. For this smaller value of we must draw a
smaller beige bar around y = 1, which in turn requires a different blue area of values
of x for which the corresponding values of f (x) lie within the -bar, as shown in the
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
96
Chapter 1
November 21, 2012
Limits
preceding graph at the right. To find the leftmost point x = a of the blue area, we solve
f (a) = 1.1:
f (a) = 1.1 =⇒
a+1
= 1.1 =⇒ a + 1 = 1.1a =⇒ 1 = 0.1a =⇒ a = 10.
a
Thus for x ∈ (10, ∞) we can guarantee that
EXAMPLE 3
x+1
x
∈ (0.9, 1.1).
A real–world example of finding δ given Fuel efficiency depends on driving speed. A typical car runs at 100% fuel efficiency when
driven at 55 miles per hour. Suppose that the fuel efficiency percentage at speed s (in MPH)
is given by E(s) = −0.033058(s 2 − 110s). If you want your car to run with at least 95% fuel
efficiency, how close to 55 miles per hour do you have to drive?
SOLUTION
In the language of limits, we are asking: For the limit lim E(s) = 100, if = 5, what is δ?
s→55
The corresponding -bar and δ-bar are shown below in the following graph of E(s):
Staying within 5% fuel efficiency
E
105
95
110
100
90
80
70
60
50
40
42.7
67.3
10 20 30 40 50 60 70 80 90100
55
s
The leftmost and rightmost values x = a and x = b of the blue δ-interval shown can be
found by using the quadratic formula to find the two solutions of the equation E(s) = 95
or by using a calculator graph to approximate values. In either case, we find a ≈ 42.7 and
b ≈ 67.3. Therefore, according to the graph you can drive anywhere between 42.7 and 67.3
miles per hour and get at least 95% fuel efficiency.
TEST YOUR
? UNDERSTANDING
When discussing limits as x→c, why do we consider punctured intervals around
x = c?
In the definition of limit, why do we need the statement to be true for all values
> 0?
In the proof of Theorem 1.6 we had L = M + k and =
(L − , L + ) and (M − , M + ) do not overlap?
k
.
2
Why does this mean that
How are left limits and right limits related to two-sided limits?
How is a limit at infinity different from an infinite limit?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.2
97
Formal Definition of Limit
EXERCISES 1.2
Thinking Back
Logical quantifiers: Determine whether each of the following statements about real numbers is true or false, and
why.
For all a, there exists some b such that b = a 2 .
For all a, there exists some b such that a = b 2 .
For all a, there exists some b such that b = a + 1.
For all integers a, there exists some integer b such that
if x ≥ a, then x > b.
For all integers a, there exists some integer b such that
if x > a, then x = b.
Solving function equations: Solve each of the following equations for x, and illustrate these solutions on a graph of y = f (x).
If f (x) = x 3 , solve f (x) = 7.5 and f (x) = 8.5.
√
If f (x) = x − 1, solve f (x) = 1.8 and f (x) = 2.2.
If f (x) = −0.033058(x 2 − 110x), solve f (x) = 90.
If f (x) = 2 − x 2 and x > 0, solve f (x) = −7.01 and
f (x) = −6.99.
If f (x) =
f (x) = 0.
x 2 − 2x − 3
and x < 0, solve f (x) = −2 and
x−1
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: For lim f (x) to be defined, the function f
x→c
must be defined at x = c.
(b) True or False: We can calculate a limit of the form
lim f (x) simply by finding f (c).
x→c
(c) True or False: If lim f (x) = 10, then f (c) = 10.
x→c
(d) True or False: If f (c) = 10, then lim f (x) = 10.
x→c
(e) True or False: A function can approach more than one
limit as x approaches c.
(f) True or False: If lim f (x) = 10, then we can make f (x)
x→4
as close to 4 as we like by choosing values of x sufficiently close to 10.
(g) True or False: If lim f (x) = ∞, then we can make f (x)
x→6
as large as we like by choosing values of x sufficiently
close to 6.
(h) True or False: If lim f (x) = 100, then we can find valx→∞
ues of f (x) between 99.9 and 100.1 by choosing values
of x that are sufficiently large.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function f and a value c such that lim f (x) happens
x→c
to be equal to f (c).
(b) A function f and a value c such that lim f (x) is not
x→c
equal to f (c).
(c) A function f and a value c such that lim f (x) exists but
x→c
f (c) does not exist.
3. What are punctured intervals, and why do we need to use
them when discussing limits?
4. Describe the punctured interval around x = 2 that has a
radius of 3 and the punctured interval around x = 4 that
has a radius of 0.25.
5. Find punctured intervals on which the function
f (x) =
1
is defined, centered around
x2 − x
(a) x = 1.5
(b) x = 0.25
(c) x = 1
6. Find punctured intervals on which the function
f (x) =
1
is defined, centered around
xln(x + 2)
(a) x = 0
(b) x = −1
(c) x = −1.5
Use interval notation to fill in the blanks that follow. Your
answers will involve δ, , N, and/or M.
7. If lim f (x) = 5, then for all > 0, there is some δ > 0
x→2
, then f (x) ∈
.
such that if x ∈
8. If lim− f (x) = 1, then for all > 0, there is some δ > 0
x→3
such that if x ∈
, then f (x) ∈
.
9. If lim f (x) = 2, then for all > 0, there is some N > 0
x→∞
, then f (x) ∈
.
such that if x ∈
10. If lim f (x) = −∞, then for all M > 0, there is some
x→∞
N > 0 such that if x ∈
, then f (x) ∈
.
11. If lim+ f (x) = ∞, then for all M > 0, there is some δ > 0
x→1
, then f (x) ∈
.
such that if x ∈
12. Sketch a labeled graph that illustrates what is going on
in the proof of Theorem 1.6 in the reading. Your graph
should include two different -bars and a graphical reason that they cannot overlap.
13. Sketch a labeled graph that illustrates what is going
on in the proof of Theorem 1.8 in the reading. Your
graph should include two different δ-bars and a graphical reason why they combine to make a punctured deltainterval.
14. Suppose f is a function with f (2) = 5 where for all > 0,
there is some δ > 0 such that if x ∈ (2 − δ, 2) ∪ (2, 2 + δ),
then f (x) ∈ (3 − , 3 + ). Sketch a possible graph of f .
15. If x ∈ (1.5, 2.5), what is the largest interval I = (4 − ,
4 + ) for which we can guarantee that x 2 ∈ I?
16. It is false that lim(x + 1.01) = 2. Express this fact in a
x→1
mathematical sentence involving δ and , to show how
the formal definition of limit fails in this case.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
98
Chapter 1
November 21, 2012
Limits
1000
= ∞. Express this fact in a mathx→∞ x
17. It is false that lim
ematical sentence involving M and N, to show how the
formal definition of limit fails in this case.
√
18. Show that the limit as x → 2 of f (x) = x − 1.1 is not
equal to 1, by finding an > 0 for which there is no corresponding δ > 0 satisfying the formal definition of limit.
Skills
Write each limit in Exercises 19–42 as a formal statement
involving δ, , N, and/or M, and sketch a graph that illustrates the roles of these constants. In the last six exercises you
may take f to be any appropriate graph.
√
x2 − 4
19. lim x + 7 = 2
=0
20. lim
x→−3
x→2 x + 2
√
21. lim (x 3 − 2) = −3
22. lim− 1 − x = 0
x→−1
x→1
x −4
=4
x→2 x − 2
√
lim+ x = 0
2
23.
25.
lim
x→0
1
=∞
x+2
1
=∞
29. lim+ x
x→2 2 − 4
x
31. lim
= −.5
x→∞ 1 − 2x
27.
33.
lim
x→−2+
lim (x 3 + x + 1) = ∞
x→∞
x→3
1
= −∞
x
1
30. lim+
= −∞
x→0 1 − e x
1
32. lim 2
=0
x→−∞ x + 1
34. lim (1 − 3x) = ∞
lim
x→0−
x→−∞
35. lim
36. lim
37.
(2 + h) − 4
=4
h
lim+ f (x) = −∞
38.
lim f (x) = L
40.
lim f (x) = ∞
42.
h→0
39.
41.
x→c
x→−∞
x→−∞
x→0
51. lim(2 − x 2 ) = −7, = 0.01
x→3
lim− (4 − x 2 ) = −5
2
1 − cos x
1
= 0, =
x
2
1 − cos x
1
50. lim
= 0, =
x→0
x
4
49. lim
52. lim(2 − x 2 ) = −7, = 0.001
x→1
28.
x→π
x→3
24. lim(x 2 − 3) = −2
26.
1
2
48. lim sin x = 0, =
1
2+h
−
1
2
=−
h
lim− f (x) = ∞
h→0
1
4
x→c
x 2 − 2x − 3
= −4, = 1
x→−1
x+1
x 2 − 2x − 3
54. lim
= −4, = 0.1
x→−1
x+1
53.
For each limit in Exercises 55–64, use graphs and algebra
to approximate the largest δ or smallest-magnitude N that
corresponds to the given value of or M, according to the
appropriate formal limit definition.
55.
56.
lim f (x) = −∞
x→−∞
lim f (x) = −∞
57.
x→∞
For each limit lim f (x) = L in Exercises 43–54, use graphs
58.
and algebra to approximate the largest value of δ such that if
x ∈ (c − δ, c) ∪ (c, c + δ), then f (x) ∈ (L − , L + ).
59.
43. lim x = 8, = 0.5
60.
44. lim x 3 = 8, = 0.25
x→2
√
45. lim x − 1 = 2, = 1
x→5
√
46. lim x − 1 = 2, = 0.2
x→5
√
2
47. lim sin x = 0, =
x→π
2
61.
x→c
3
lim
x→2
62.
63.
64.
1
= ∞, M = 1000, find largest δ > 0
x2 − 1
1
lim
= ∞, M = 10,000, find largest δ > 0
x→1+ x 2 − 1
3x
lim
= 3, = 0.5, find smallest N > 0
x→∞ x + 1
3x
= 3, = 0.1, find smallest N > 0
lim
x→∞ x + 1
lim ln x = ∞, M = 100, find smallest N > 0
lim
x→1+
x→∞
lim ln x = ∞, M = 100,000, find smallest N > 0
x→∞
1
2
lim 3 x = 0, = , find smallest-magnitude N < 0
x→−∞
1
4
lim 3 x = 0, = , find smallest-magnitude N < 0
x→−∞
lim (4 − x 2 ) = −∞, M = −100, find smallest N > 0
x→∞
lim (4 − x 2 ) = − ∞, M = − 10,000, find smallest N > 0
x→∞
Applications
65. Every month, Jack hides $50 under a broken floorboard to
save up for a new boat. After t months of saving, he will
have F(t) = 50t dollars.
(a) The boat Jack wants costs at least $7,465. How many
months does Jack have to save money before he will
have enough to pay for the boat? Illustrate this information on a graph of F(t).
(b) Suppose a different boat costs M dollars. Will there
be a value t = N for which F(N) > M? What does
this mean in real-world terms? Illustrate the roles of
M and N on a graph of F(t).
Money saved for Jack’s boat
F
10,000
7,500
5,000
2,500
48
96
CONFIRMING PAGES
144
192
t
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.2
66. Len’s company produces different-sized cylindrical cans
that are each 6 inches tall. The cost to produce a can with
radius r is C(r) = 10π r2 + 24π r cents.
Different cans with radius 6 inches
r
Cost of producing a can
C
67. You work for a company that sells velvet Elvis paintings.
The function N( p) = 9.2p2 − 725p + 16, 333 predicts the
number N of velvet Elvis paintings that your company will
sell if they are priced at p dollars each, and is shown in the
following graph. The Presley estate does not allow you to
charge more than $50 per painting.
500
Velvet Elvis paintings sold
N
400
6 in.
99
Formal Definition of Limit
300
10,000
200
8,000
100
0.5 1.0 1.5 2.0 2.5 3.0
6,000
r
4,000
2,000
(a) Len’s boss wants him to construct the cans so that
the cost of each can is within 25 cents of $4.00. Given
these cost requirements, what is the acceptable range
of values for r?
(b) Len’s boss now says that he wants the cans to cost
within 10 cents of $4.00. Under these new cost requirements, what is the acceptable range of values for r?
(c) Interpret this problem in terms of δ and ranges.
Specifically, what is c? What is L? What is for part
(a) and part (b)? What are the corresponding values
of δ? Illustrate these values of c, L, , and δ on a graph
of C(r).
10
20
30
40
50
p
(a) Use a graphing utility to estimate the price your company should charge per painting if it wishes to sell
6000 velvet Elvis paintings.
(b) Find the range of prices that would enable your
company to sell between 5000 and 7000 velvet Elvis
paintings.
(c) Interpret this problem in terms of δ and ranges.
Specifically, what is c? What is L? What is ? What
is the corresponding value of δ? Illustrate these values of c, L, , and δ on a graph of N( p).
Proofs
Prove the four limit statements in Exercises 68–71. In the next
section we will present a systematic method for such proofs.
68. Prove that lim 3x = 3, with these steps:
x→1
(a) What is the δ– statement that must be shown to
prove that lim 3x = 3?
x→1
(b) Argue that x ∈ (1 − δ, 1) ∪ (1, 1 + δ) if and only if
−δ < x − 1 < δ, with x = 1. Then use algebra to
show that this means that 0 < |x − 1| < δ.
(c) Argue that 3x ∈ (3 − , 3 + ) if and only if − <
3(x − 1) < . Then use algebra to show that this
means that 3|x − 1| < .
(d) Given any particular > 0, what value of δ would
guarantee that if 0 < |x − 1| < δ, then 3|x − 1| < ?
Your answer will depend on .
(e) Put the previous four parts together to prove the limit
statement.
69. Prove that lim(7 − x) = 5, with these steps:
x→2
(a) What is the δ– statement that must be shown to
prove that lim 7 − x = 5?
x→2
(b) Argue that x ∈ (2 − δ, 2) ∪ (2, 2 + δ) if and only if
−δ < x − 2 < δ. Then use algebra to show that this
means that 0 < |x − 2| < δ.
(c) Argue that 7 − x ∈ (5 − , 5 + ) if and only if
− < 2 − x < . Then use algebra to show that
this means that |x − 2| < .
(d) Given any particular > 0, what value of δ would
guarantee that if 0 < |x − 2| < δ, then |x − 2| < ?
Your answer will depend on .
(e) Put the previous four parts together to prove the limit
statement.
70. Prove that lim+
x→0
1
= ∞, with these steps:
x
(a) What is the M–δ statement that must be shown to
prove that lim+
x→0
1
= ∞?
x
(b) Argue that x ∈ (0, 0 + δ) if and only if 0 < x < δ.
1
1
∈ (M, ∞) if and only if x < . You may
x
M
(c) Argue that
assume that M > 0.
(d) Given any particular M > 0, what value of δ would
1
guarantee that if 0 < x < δ, then x <
? Your
M
answer will depend on M.
(e) Put the previous four parts together to prove the limit
statement.
71. Prove that lim
x→∞
1
= 0, with these steps:
x
(a) What is the –N statement that must be shown to
prove that lim
x→∞
1
= 0?
x
(b) Argue that x ∈ (N, ∞) if and only if x > N.
(c) Argue that
1
1
∈ (0−, 0+) if and only if − < < .
x
x
Then argue that for this limit, it suffices to consider
1
0 < < .
x
(d) Given any particular , what value of N > 0 would
1
< ? Your
guarantee that if x > N, then 0 <
x
answer will depend on .
(e) Put the previous four parts together to prove the limit
statement.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
100
Chapter 1
November 21, 2012
Limits
Thinking Forward
Continuity: As you have already seen, sometimes lim f (x) is
x→c
equal to f (c), and sometimes it is not. As we will see in
Section 1.4, when the limit of a function f as x → c does
happen to be equal to the value of f (x) at x = c, we say that
the function f is continuous at x = c.
1.3
You may have heard the following loose, only partially
accurate definition of continuity in a previous class:
A function is continuous if you “can draw it without
picking up your pencil.” Why does it make sense that
this would be related to the definition just presented
of continuity in terms of limits?
State the limit-definition of continuity with a formal
δ– statement.
The function f (x) = x 2 is continuous at every point.
Use this fact and the formal definition of continuity to
calculate lim x 2 , lim x 2 , and lim x 2 .
x→2
x→5
x→−4
Limits of Sequences: We say that an infinite sequence of real
numbers a 1 , a 2 , a 3 , . . . , a k , . . . converges to a limit L, and we
write lim s k = L, if for all > 0, there exists some N > 0
k→∞
such that if k > N, then |a k − L| < .
Use algebra to solve the inequality |a k − L| < for a k .
, where
Your answer should be in the form a k ∈
the blank is filled in with interval notation.
Relate the definition of the convergence of a sequence
to the definition of a limit at infinity.
DELTA-EPSILON PROOFS*
Developing an equivalent algebraic definition of limits from our geometric definition
Finding delta in terms of epsilon so that we can prove a limit statement
The formal logic of writing delta–epsilon proofs
Describing Limits with Absolute Value Inequalities
In Definition 1.5, we formally defined the limit statement lim f (x) = L to mean that for all
x→c
> 0, there exists δ > 0 such that whenever x ∈ (c − δ, c) ∪ (c, c + δ), we can guarantee that
f (x) ∈ (L − , L + ). This definition of limit has a very geometric flavor, since it is stated
in the language of -intervals and punctured δ-intervals. That kind of language is useful
when looking at specific values of or δ, but not as useful when trying to prove that every
value of has a corresponding value of δ. For the purposes of proving limit statements, we
give the following algebraic definition of limit and prove that it is equivalent to our previous
geometric definition:
DEFINITION 1.10
Algebraic Definition of Limit
The limit lim f (x) = L means that for all > 0, there exists δ > 0 such that
x→c
if 0 < |x − c| < δ, then | f (x) − L| < .
THEOREM 1.11
Equivalence of Geometric and Algebraic Definitions of Limit
The two definitions of limit in Definition 1.5 and 1.10 are equivalent. Specifically,
(a) x ∈ (c − δ, c) ∪ (c, c + δ) if and only if 0 < |x − c| < δ;
(b) f (x) ∈ (L − , L + ) if and only if | f (x) − L| < .
Proof. We will begin by proving part (b), since it is slightly easier. The statement f (x) ∈ (L − ,
L + ) means that L − < f (x) < L + . Subtracting L from all three parts of this double inequality,
we get − < f (x) − L < . This is precisely the solution of the inequality | f (x) − L| < .
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.3
101
Delta-Epsilon Proofs*
For part (a) we have a similar situation, except that we must deal with a punctured interval.
The statement x ∈ (c − δ, c) ∪ (c, c + δ) means that c − δ < x < c + δ and x = c. Subtracting c from
all three parts of the double inequality, we get −δ < x − c < δ, which is the solution set for the
inequality |x − c| < δ. The fact that x = c means that x − c = 0. This is equivalent to saying that
|x − c| = 0, and since the absolute value of a number is always positive or zero, it is also equivalent
to saying that |x − c| > 0. Therefore x ∈ (c − δ, c) ∪ (c, c + δ) if and only if 0 < |x − c| < δ.
Finding a Delta for Every Epsilon
With our new algebraic definition of limit, we have the final tool we need to be able to
effectively prove limit statements. This is the first step towards being able to calculate limits,
something that, perhaps surprisingly, we do not yet know how to do.
Consider for example the limit statement lim (3x − 1) = 5. We can examine this limit
x→2
with a table of values, noticing that as the values of x approach x = 2 from the left and the
right, the corresponding values of 3x − 1 approach 5:
x
1.9
1.99
1.999
2
2.001
2.01
2.1
3x − 1
4.7
4.97
4.997
-
5.003
5.03
5.3
We can also investigate the limit statement lim (3x − 1) = 5 with the following graph of
x→2
f (x) = 3x − 1 at the left, noticing that a sequence of values of x approaching x = 2 from
either the left or the right determines a sequence of values of f (x) that approach y = 5:
1
1
,
If x ∈ 2 − , 2 ∪ 2, 2 +
As x → 2, f (x) → 5
3
3
then f (x) ∈ (5 − 1, 5 + 1)
y
1
1
If x ∈ 2 − , 2 ∪ 2, 2 +
,
6
6
1
1
then f (x) ∈ 5 − , 5 +
y
8
8
8
7
7
7
6
6
6
5
5
5
4
4
4
3
3
3
2
2
2
1
1
1
2
3
2
y
2
1
x
1
3
2
x
1
2
3
x
We can even show that the formal geometric definition of limit holds for a particular , by
finding some δ for which x ∈ (2 − δ, 2) ∪ (2, 2 + δ) guarantees that 3x − 1 ∈ (5 − , 5 + ).
1
For example, it turns out that if = 1, then δ = will work, as illustrated in the middle
1
2
figure shown, and if = , then δ =
1
6
3
will work, as in the rightmost figure.
Although the preceding investigations are encouraging evidence, none of them prove
that lim (3x − 1) is equal to 5. In the table, perhaps we did not consider values of x close
x→2
enough to 2 to see the true behavior of f (x) = 3x − 1, or perhaps the limit is actually 5.0001
and not 5. The three graphs shown each have the same possible problems. To prove that
lim (3x−1) = 5 we would need to know for certain that we could get values of f (x) arbitrarily
x→2
close to y = 5 by choosing values of x that are sufficiently close to x = 2. For every epsilon
we would have to be able to find a delta.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
102
Chapter 1
November 21, 2012
Limits
Specifically, in this example, we need to show that for any choice of > 0, we can find
some δ > 0 such that whenever a real number x satisfies 0 < |x − 2| < δ, we can also
guarantee that x satisfies |(3x − 1) − 5| < . Notice that |(3x − 1) − 5| = |3x − 6| = 3|x − 2|,
so what we really need is to conclude that 3|x − 2| < , i.e., that |x − 2| < . But that is easy
3
if we are allowed to choose a different δ for each : Given any > 0, simply choose δ = .
3
Then whenever 0 < |x −2| < δ, we also have |x −2| < , as desired. Notice that in both the
3
second and third of the graphs shown previously, we did in fact choose δ to be one-third
of . We have just proved the limit statement lim (3x − 1) = 5 by showing that every value
of has a corresponding value of δ =
holds.
3
x→2
for which the formal algebraic definition of limit
Writing Delta-Epsilon Proofs
We have just proved that lim (3x − 1) = 5, but our proof meandered about in a paragraph
x→2
of discussion. We now present a systematic way to write up delta-epsilon proofs for limit
statements. Remember that in our example what we must show is the following doubly
quantified logical implication:
“For all > 0, there exists a δ > 0 such that if 0 < |x − 2| < δ, then |(3x − 1) − 5| < .”
Logically, to prove such a statement we must first let be an arbitrary positive number and
then choose a value for δ in terms of . We must then show that for all values of x with
0 < |x − 2| < δ, we can also say that |(3x − 1) − 5| < .
In our example, we already know that for any > 0 we should choose δ =
this fact, we could arrange our proof very concisely as follows:
Proof. Given > 0, choose δ =
3
. Using
. For all x with 0 < |x − 2| < δ, we have
= .
|(3x − 1) − 5| = |3x − 6| = |3(x − 2)| = 3|x − 2| < 3δ = 3
3
3
Therefore whenever 0 < |x − 2| < δ, we also have |(3x − 1) − 5| < .
The first three equals signs in the proof are simple algebra. The less-than step followed
from our assumption that |x − 2| < δ. The starred equality followed from the fact that
δ= .
3
In general we will not know at the outset what to choose for δ, as we did just now. In
those cases we can either do a side-calculation to find δ in advance or just leave a blank
space for δ and continue with the proof. When we get to the starred equality, after using
the assumption 0 < |x − c| < δ, we will be able to see what to choose for δ and can fill in
the blank as if we knew it all along; see Example 2. Throughout the examples we will try
our hand at proving all kinds of limit statements, including one-sided limits, infinite limits,
and limits at infinity.
Examples and Explorations
EXAMPLE 1
Finding delta as a function of epsilon
Find formulas for δ in terms of for each of the following limit statements:
(a) lim (2x + 1) = 9
x→4
(b) lim (x 2 − 4x + 5) = 1
x→2
Then use those formulas to find punctured δ-intervals for = 1 and = 0.5.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.3
103
Delta-Epsilon Proofs*
SOLUTION
(a) The limit statement lim (2x + 1) = 9 means that given any > 0, there is some δ > 0
x→4
so that if 0 < |x − 4| < δ, then we can conclude that |(2x + 1) − 9| < . We have
|(2x + 1) − 9| = |2x − 8| = |2(x − 4)| = 2|x − 4|,
so |(2x + 1) − 9| < when 2|x − 4| < , i.e., when |x − 4| <
should choose δ = . In particular, when = 1, we have δ =
2
δ-interval (3.5, 4) ∪ (4, 4.5). When = 0.5, we choose δ =
the punctured δ-interval (3.75, 4) ∪ (4, 4.25).
0.5
2
1
2
2
. Therefore we
= 0.5 and punctured
= 0.25, which gives us
(b) The limit statement lim (x 2 − 4x + 5) = 1 means that for any > 0, we can find δ > 0
x→2
so that whenever 0 < |x − 2| < δ, we can conclude that |(x 2 − 4x + 5) − 1| < . Notice
that
|(x 2 − 4x + 5) − 1| = |x 2 − 4x + 4| = |(x − 2)2 | = |x − 2|2 ,
which is clearly very closely related to our δ-inequality 0 < √
|x − 2| < δ. In fact, the
inequality |x−2|2 < is equivalent to the inequality |x−2| < . Therefore
we should
√
√
= 1 and thus
choose δ = . In particular, when = 1, we should choose δ = 1 √
punctured δ-interval (1, 2) ∪ (2, 3). When = 0.5, we should choose δ = 0.5 ≈ 0.707
and thus punctured δ-interval (1.293, 2.707).
CHECKING
THE ANSWER
To check that our formulas for δ in terms of are reasonable, we can graph the functions
and the punctured δ-intervals that we found in each case. For the limit lim (2x + 1) = 9,
x→4
δ = 0.5 looks right for = 1 and δ = 0.25 looks right for = 0.5:
If x ∈ (4 − 0.5, 4) ∪ (4, 4 + 0.5),
then f (x) ∈ (9 − 1, 9 + 1)
If x ∈ (4 − 0.25, 4) ∪ (4, 4 + 0.25),
then f (x) ∈ (9 − 0.5, 9 + 0.5)
y
y
11
10
9
8
7
6
5
4
3
2
1
1
2
3
4
5
x
11
10
9
8
7
6
5
4
3
2
1
1
2
3
4
5
x
For lim (x 2 − 4x + 5) = 1, δ = 1 looks right for = 1 and δ ≈ 0.707 looks right for = 0.5:
x→2
If x ∈ (2 − 1, 2) ∪ (2, 2 + 1),
then f (x) ∈ (1 − 1, 1 + 1)
If x ∈ (2 − 0.707, 2) ∪ (2, 2 + 0.707),
then f (x) ∈ (1 − 0.5, 1 + 0.5)
y
y
4
4
3
3
2
2
1
1
1
2
3
4
x
1
2
CONFIRMING PAGES
3
4
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
104
Chapter 1
EXAMPLE 2
November 21, 2012
Limits
Writing basic delta–epsilon proofs
Write delta–epsilon proofs for each of the following limit statements:
(b) lim (x 2 − 4x + 5) = 1
(a) lim (2x + 1) = 9
x→2
x→4
SOLUTION
(a) To show that lim (2x + 1) = 9 we must start with an arbitrary > 0, choose some
x→4
δ > 0, and then show that for all x with 0 < |x − 4| < δ, we also have |(2x + 1) − 9| < .
From the previous example we know that we should choose δ = .
2
Proof. Given > 0, choose δ =
. For all x with 0 < |x − 4| < δ, we have
= .
|(2x + 1) − 9| = |2x − 8| = |2(x − 4)| = 2|x − 4| < 2δ = 2
2
2
Therefore whenever 0 < |x − 4| < δ, we also have |(2x + 1) − 9| < .
If we had not already known that we should choose δ = , we could have determined
2
that choice when we reached the starred equality. At that point we had shown that
|(2x + 1) − 9| was less than 2δ. What we were trying to show was that |(2x + 1) − 9|
was less than . If 2δ were equal to , then we would be done; solving 2δ = for δ, we
arrive at the choice δ = .
2
(b) To show that lim (x 2 − 4x + 5) = 1 we must start with an arbitrary > 0, choose some
x→2
δ > 0, and then show that for all x with 0 < |x − 2| < δ, we also have |(x 2 −√4x +
5) − 1| < . From the previous example we know that we should choose δ = .
√
. For all x with 0 < |x − 2| < δ, we have
√
2
2
|(x − 4x + 5) − 1| = |x − 4x + 4| = |(x − 2)2 | = |x − 2|2 < δ 2 = ( )2 = .
Proof. Given > 0, choose δ =
Thus we can conclude that whenever 0 < |x − 2| < δ, we also have |(x 2 − 4x + 5) − 1| < .
√
Again, if we had not already known that we should choose δ = , then at the starred
2
2
equality we would
√ have δ where we wish to have . Solving δ = for δ, we get our
choice of δ = .
EXAMPLE 3
Proofs for one-sided and infinite limits
Give formal proofs for each of the following limit statements:
√
1
(a) lim+ x − 1 = 0
(c) lim−
(b) lim = 0
x→1
x→∞ x
x→3
1
=∞
3−x
SOLUTION
√
x − 1 = 0 means that for all > 0, there exists δ > 0 such
x→1
√
that if x ∈ (1, 1 + δ), then | x − 1 − 0| < . Other than a small bit of extra work to
translate the meaning of x ∈ (1, 1 + δ), this is a standard delta-epsilon proof.
(a) The limit statement lim+
Proof. Given > 0, choose δ = 2 . (The reason for that choice of δ will become clear at the
starred equality that follows.) Suppose x ∈ (1, 1 + δ). Then 1 < x < 1 + δ, which means that
0 < x − 1 < δ. We then have
√
√
√
√
√
| x − 1 − 0| = | x − 1| = x − 1 < δ = 2 = .
√
Thus we can conclude that whenever x ∈ (1, 1 + δ), we also have | x − 1 − 0| < .
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.3
(b) Geometrically, the limit statement lim
N > 0 such that if x ∈ (N, ∞),
1
Delta-Epsilon Proofs*
105
= 0 means that for all > 0, there exists
x→∞ x
1
then ∈ (0
x
− , 0 + ). Algebraically, the implication
after the quantifiers can be rewritten as follows: If x > N, then
what we shall prove.
1
x
− 0 < . This is
1
. (The reason for that choice of N will become clear at
1
1
the starred equality that follows.) If x > N, then < and x is positive, and therefore
x
N
Proof. Given > 0, choose N =
1
1
1
1
1
= <
−0 =
=
= .
1
x
x
x
N
Thus we can conclude that whenever x > N, we also have
(c) Geometrically, the limit lim−
x→3
1
3−x
such that if x ∈ (3 − δ, 3), then
1
− 0 < .
x
= ∞ means that for all M > 0, there exists δ > 0
1
3−x
∈ (M, ∞). Note that x ∈ (3 − δ, 3) means that
3 − δ < x < 3 and therefore that −δ < x − 3 < 0. Multiplying by −1 and flipping
inequalities, this becomes δ > −(x − 3) > 0, or equivalently, 0 < 3 − x < δ. Hence the
implication in our limit statement can be expressed as follows: If 0 < 3 − x < δ, then
1
3−x
> M. This is what we will prove:
1
. (As usual, the reason for this choice will become clear
M
1
1
at the starred equality that follows.) For all x with 0 < 3 − x < δ, we have
> , and
3−x
δ
Proof. Given M > 0, choose δ =
therefore
1
1
1
> =
= M.
1
3−x
δ
M
Thus we can conclude that whenever 0 < 3 − x < δ, we also have
EXAMPLE 4
1
> M.
3−x
A delta–epsilon proof where it is necessary to bound delta from above
Write a delta-epsilon proof for the limit statement lim 5x 4 = 80.
x→2
SOLUTION
To prove the limit statement lim 5x 4 = 80, we must show that for all > 0, there exists a
x→2
choice of δ > 0 such that whenever 0 < |x − 2| < δ, we also have |5x 4 − 80| < . In our
previous delta-epsilon proofs, the algebra has always magically worked out nicely. In this
example there will be a point at which we get stuck. What will get us out of the jam will
be assuming that δ is no larger than 1. This assumption will allow us to put a bound on
an expression that would otherwise be in our way. For that reason, our choice of δ in this
proof will be the minimum of 1 and an expression that depends on epsilon.
, i.e., the smaller of 1 and
Proof. Given > 0, choose δ = min 1,
. (The reason for this
325
325
elaborate choice of δ will be made clear at the two starred inequalities that follow.) If x is a real
number with 0 < |x − 2| < δ, then we have
|5x 4 − 80| = |5(x 4 − 16)| = 5|(x 2 − 4)(x 2 + 4)|
= 5|(x − 2)(x + 2)(x 2 + 4)|
= 5|x − 2| · |(x + 2)(x 2 + 4)|
< 5δ · |(x + 2)(x 2 + 4)|.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
106
Chapter 1
November 21, 2012
Limits
Because our choice of δ ensures that δ is at most 1, we can say that 0 < |x−2| ≤ 1 and therefore that
x is between 1 and 3. This means that we can bound the troublesome quantity |(x + 2)(x 2 + 4)| from
above; it is largest when x = 3, so we know that |(x + 2)(x 2 + 4)| is at most |(3 + 2)(32 +
4)| =65.
at
Combining this result with the work we just did, we have (using our choice of δ = min 1,
325
the two starred inequalities)
|5x 4 − 80| < 5δ · 65 = 325 δ ≤ 325
325
= .
Thus we can conclude that whenever 0 < |x − 2| < δ, we also have |5x 4 − 80| < .
TEST YOUR
? UNDERSTANDING
Why do we say that Definition 1.5 from the previous section is geometric in nature
while we say that Definition 1.10 is algebraic? In what situations might one definition
be more useful than another?
How can we express the formal definitions for one-sided limits, infinite limits, and
limits at infinity “algebraically,” in the spirit of Definition 1.10?
Suppose that we have a limit of the form lim f (x) = L and that we can find a value of δ
x→c
for some given value of . Why does this not prove definitively that lim f (x) = L?
x→c
In a delta-epsilon proof, how do we come up with a choice for δ in terms of ?
Why is it sometimes necessary to require that δ ≤ 1 in a delta-epsilon proof?
EXERCISES 1.3
Thinking Back
Inequalities: Find the solution sets of each of the following
inequalities.
Logical implications: Suppose A and B are statements and that
the implication “If A, then B” holds.
0 < |x − 2| < 0.5
0 < |x + 5| < 0.1
|x 2 − 4| < 0.5
|(3x + 1) − 10| < 1
1
< 0.01
x2
1
> 1000
x2
If A is true, then what, if anything, can you conclude
about B, and why?
If A is false, then what, if anything, can you conclude
about B, and why?
If B is true, then what, if anything, can you conclude
about A, and why?
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If x = c, then |x − c| is strictly greater
than zero.
(b) True or False: If |x −c| is strictly greater than zero, then
x = c.
(c) True or False: x is a solution of 0 < |x − c| < δ if and
only if c − δ < x < c + δ.
(d) True or False: If 0 < |x − c| < δ, then x ∈ (c − δ, c) ∪
(c, c + δ).
(e) True or False: If | f (x)−L| < , then L− < f (x) < L+.
(f) True or False: If f (x) ∈ (L − , L + ), then 0 < f (x) <
|L + |.
(g) True or False: The fact that 0 < |x − 3| < 0.25
guarantees that |(2x − 1) − 5| < 0.5 proves that
lim(2x − 1) = 5.
x→3
(h) True or False: lim(2x − 1) = 5 means that for all δ > 0
x→3
there is some > 0 such that if 0 < |x − c| < δ, then
|(2x − 1) − 5| < .
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.3
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function f with values given in the following table
but whose limit as x → 2 is not equal to 5:
x
1.9 1.99 1.999 2 2.001 2.01 2.1
f (x) 4.7 4.97 4.997 - 5.003 5.03 5.3
(b) An inequality involving absolute values whose solution set is (2.75, 3) ∪ (3, 3.25).
(c) An inequality involving absolute values whose solution set is (−0.01, 0) ∪ (0, 0.01).
Determine whether each implication that follows is true or
false. Use graphs to justify any implications that are true, and
counterexamples for any implications that are false.
13.
14.
15.
16.
17.
lim
3
= −3.
x+1
If 0 < |x − 2| < 1, then |x 2 − 4| < 0.5.
If 0 < |x − 2| < 0.2, then |x 2 − 4| < 1.
If 0 < |x − 0| < 0.75, then |x 2 − 0| < 0.5.
If 0 < |x + 2| < 0.1 then |x 2 − 4| < 0.4.
If 0 < |x + 2| < 0.075, then |x 2 − 4| < 0.4.
18. In Example 2 we proved that lim(x 2 − 4x + 5) = 1. Use
x→2
the proof to find values of δ corresponding to (a) = 1,
(b) = 0.1, and (c) = 0.01. Illustrate that your choices
of δ work by examining a graph of f (x) = x 2 − 4x + 5 and
sketching appropriate and δ intervals.
3. Write down the formal delta-epsilon statement you
would have to prove in order to prove the limit statement
x→−2
107
Delta-Epsilon Proofs*
19. In Example 4 we proved that lim 5x 4 = 80. Use the
x→2
4. Suppose you show that |(1 − 2x) − (−5)| < 0.05 for all x
with 0 < |x − 3| < 0.025. Explain why this does not prove
that lim(1 − 2x) = −5.
x→3
5. Write down a mathematical equation that expresses the
sentence “x is not equal to 5, and the distance between
x and 5 is less than 0.01.” Then write an equation that
means “the distance between f (x) and −2 is less than
0.5.”
6. Why do we have 0 < |x − c| < δ instead of just |x − c| < δ
in Definition 1.10?
Write each of the following inequalities in interval notation:
proof to find values of δ corresponding to (a) = 5,
(b) = 0.01, and (c) = 350. Illustrate that your choices
of δ work by examining a graph of f (x) = 5x 4 and sketching appropriate and δ intervals.
20. Use algebra to solve the inequality 0 < |x − c| < δ and
show that its solution set is x ∈ (c − δ, c) ∪ (c, c + δ).
21. Use algebra to solve the inequality | f (x) − L| < and
show that its solution set is f (x) ∈ (L − , L + ).
22. Suppose f (x) = mx + b is a linear function with m = 0,
and let c be any real number.
, then
(a) Show that for all > 0, if 0 < |x − c| <
|m|
| f (x) − f (c)| < .
(b) What does the implication in part (a) have to do with
limits?
(c) Illustrate the implication in part (a) with a labeled
graph. Explain in terms of slopes why it makes sense
.
that given > 0, the corresponding δ > 0 is δ =
8. 0 < |x + 3| < 0.05
7. 0 < |x − 2| < 0.1
10. |(3x + 1) − 2| < 0.1
9. |(x 2 − 1) + 3| < 0.5
12. 0 < |x − c| < δ
11. | f (x) − L| < |m|
Skills
Use algebra to find the largest possible value of δ or smallest possible value of N that makes each implication in Exercises 23–28 true. Then verify and support your answers with
labeled graphs.
23. If 0 < |x − 2| < δ, then |(3x − 1) − 5| < 0.25.
1
1
< 0.2.
−
24. If 0 < |x − 3| < δ, then
x
3
√
25. If x ∈ (1, 1 + δ), then x − 1 − 0 < 0.5.
26. If x ∈ (3 − δ, 3), then
27. If x > N, then
1
> 1000.
3−x
1
− 0 < 0.001.
x2
28. If x > N, then 1 − 2x < −500.
For each limit statement lim f (x) = L in Exercises 29–40, use
x→c
algebra to find δ > 0 in terms of > 0 so that if 0 <
|x − c| < δ, then | f (x) − L| < .
29. lim(x + 5) = 8
30.
31. lim(3 − 4x) = 3
32. lim(3x + 8) = 11
33. lim(5x 2 − 1) = −1
34. lim(x 2 − 6x + 5) = −4
x→3
x→0
x→0
lim (4 − 2x) = 8
x→−2
35. lim(x 2 − 4x + 6) = 2
x→2
36. lim(x 3 + 1) = 1
x→0
1
1
37. lim = ; you may assume δ ≤ 1
x→2 x
2
1
1
38. lim = ; you may assume δ ≤ 1
x→3 x
3
39. lim(x 2 − 2x − 3) = 0; you may assume δ ≤ 1
x→3
40. lim 2x 4 = 2; you may assume δ ≤ 1
x→1
For each limit statement in Exercises 41–44, use algebra to
find δ or N in terms of or M, according to the appropriate
formal limit definition.
√
41. lim + (1 + x + 2 ) = 1, find δ in terms of x→−2
x−1
= 1, find N in terms of x
1
43. lim−
= ∞, find δ in terms of M
x→1 1 − x
42.
44.
lim
x→∞
lim (x 2 + 2) = ∞, find N in terms of M
x→∞
x→1
x→3
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
108
Chapter 1
November 21, 2012
Limits
Applications
For Exercises 45 and 46, suppose you work for a company that
manufactures gourmet soup cans. The material for the curved
sides of the cans costs 0.25 cent (a quarter of a cent) per square
Cost of materials for producing a soup can
0.1 cents per
linear inch
r
r
5 in.
0.25 cents per
square inch
0.5 cents per
square inch
inch, the material for the top and bottom costs 0.5 cent per
square inch, and the reinforcing weld around the seams costs
0.1 cent per linear inch. The seams run around the edges of
the top and bottom and also in a straight line from the top to
the bottom of the curved side.
45. Find a formula for the cost C(r) of producing a gourmet
soup can with radius r and height 5 inches, and answer
the following questions:
(a) What is the radius of a can that is 5 inches tall and
costs 30 cents to produce?
(b) Your manager wants you to produce 5-inch-tall cans
that cost between 20 and 40 cents. Write this requirement as an absolute value inequality.
(c) What range of radii would satisfy your manager?
Write an absolute value inequality whose solution set
lies inside this range of radii.
46. Find a formula for the cost C(h) of producing a gourmet
soup can with height h and radius 2 inches, and answer
the following questions:
(a) What is the height of a can that has radius 2 inches
and costs 45 cents to produce?
(b) Your manager wants you to produce 2-inch-radius
cans that cost between 40 and 50 cents. Write this
requirement as an absolute value inequality.
(c) What range of heights would satisfy your manager?
Write an absolute value inequality whose solution set
lies within this range of heights.
Proofs
Write delta-epsilon proofs for each of the limit statements
lim f (x) = L in Exercises 47–60.
x→c
47. lim(2x + 4) = 6
x→1
49.
lim (x + 2) = −4
x→−6
48. lim(3 − 4x) = −5
x→2
50.
lim (1 − x) = 4
x→−3
51. lim(6x − 1) = 23
52. lim(3x − 11) = 13
53. lim(3x + 1) = 1
54. lim(x 2 − 6x + 11) = 2
55. lim(2x − 4x + 3) = 1
56. lim(3x − 12x + 15) = 3
x −1
=2
x→1 x − 1
√
59. lim+ x − 5 = 0
x 2 − 3x + 2
=1
x→2
x−2
√
60. lim+ 3 2x − 4 = 0
x→4
2
x→0
2
x→1
2
57. lim
x→5
x→8
x→3
2
x→2
58. lim
x→2
For each of the limit statements in Exercises 61–66, write a
δ–M, N–, or N–M proof, according to the type of limit
statement.
1
=∞
x+2
2x − 1
63. lim
=2
x→∞
x
61.
65.
1
= −∞
x+2
2x − 1
64. lim
=2
x→−∞
x
lim
62.
x→−2+
lim (3x − 5) = ∞
66.
x→∞
lim
x→−2−
lim (3x − 5) = −∞
x→−∞
Prove each of the limit statements in Exercises 67–72. You
will have to bound δ.
68. lim (x 2 − 2x − 3) = 0
67. lim(x 2 − 2x − 3) = 0
x→3
x→−1
69. lim(x 2 − 6x + 7) = 2
70. lim(x 2 − 6x + 7) = 2
4
71. lim 2 = 1
x→2 x
72. lim
x→5
x→1
x→3
18
=2
x2
Thinking Forward
Calculating limits: We still do not have a way to calculate
limits easily. In the following problems you will develop rules
for calculating limits of some very simple functions.
x→c
for any real number c. Then use a delta–epsilon
argument to prove it. (Hint: You will need to assume that
δ ≤ 1.)
• lim x
• lim x
• lim x
2
x→0
• lim(2x−3)
Explain why it makes intuitive sense that lim x 2 = c2
x→c
• lim x
x→−1
Explain why it makes intuitive sense that lim x = c for
any real number c. Then use a delta–epsilon argument
to prove it.
Use the preceding two problems and the result of
Exercise 22 to calculate the following limits:
x→0
• lim x
x→π
x→4
2
x→5
• lim(1 − x)
x→1
• lim x 2
x→−2
• lim(3x+1)
x→3
When calculating each of these limits lim f (x), you
x→c
simply used the value of f (c). Will that method always
work for any limit? Why or why not?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
1.4
109
Continuity and Its Consequences
CONTINUITY AND ITS CONSEQUENCES
Continuity of functions at points and on intervals, and basic types of discontinuities
Simple functions that are continuous on their domains
The Extreme Value Theorem and the Intermediate Value Theorem
Defining Continuity with Limits
Intuitively, a function is continuous if its graph has no breaks, jumps, or holes. Loosely
speaking, you can sketch the graph of a continuous function “without picking up your
pencil.” We can make the notion of continuity more precise by using limits. For example,
consider the following four graphs:
y = f (x)
y = g(x)
y
y = h(x)
y
y = k(x)
y
y
4
4
4
4
3
3
3
3
2
2
2
2
1
1
1
1
1
2
3
4
x
1
2
3
4
x
1
2
3
4
x
1
2
3
4
x
While the first graph has no breaks or holes, the remaining three graphs all have some sort
of bad behavior at x = 1. It turns out that limits as x → 1 detect exactly this bad behavior:
In each case, the limit as x → 1 is not the same as the value at x = 1. For example, lim g(x)
x→1
does not exist, but g(1) = 2. For h(x), lim h(x) is equal to 1, while the value h(1) is equal to 2.
x→1
Finally, for k(x), the limit is lim k(x) = 1 but the value k(1) does not exist. On the other hand,
x→1
for f (x), we have both lim f (x) = 1 and f (1) = 1.
x→1
The preceding examples suggest the following definition: A function is continuous at
a point x = c if its limit as x → c is equal to a real number that is the same as the value of
the function at x = c.
DEFINITION 1.12
Continuity of a Function at a Point
A function f is continuous at x = c if lim f (x) = f (c).
x→c
By considering one-sided limits, we can get a more detailed picture of continuity. For
example, with the previous function g(x), the left limit as x → 1 is not equal to the value g(1)
but the right limit is. We say that g(x) is right continuous at x = 1 but not left continuous.
DEFINITION 1.13
Left and Right Continuity at a Point
A function f is left continuous at x = c if lim− f (x) = f (c) and is right continuous at
x = c if lim+ f (x) = f (c).
x→c
x→c
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
110
Chapter 1
November 21, 2012
Limits
Sometimes it is convenient to talk about continuity of a function on an interval. We say
that a function is continuous on an open interval if it is continuous at each point in the
interval. For non-open intervals we also require one-sided continuity as we approach any
closed endpoints.
Continuity of a Function on an Interval
A function f is continuous on an interval I if it is continuous at every point in the interior
of I, right continuous at any closed left endpoint, and left continuous at any closed right
endpoint.
DEFINITION 1.14
The graphs that follow provide examples of continuity on the four possible types of
bounded intervals. For example, a function f is continuous on I = (1, 3] if it is continuous at every point in the interior (1, 3) and left continuous at the right endpoint x = 3, as
shown in the third figure. In terms of limits this means that lim f (x) = f (c) for all c ∈ (1, 3)
x→c
and lim− f (x) = f (3).
x→3
continuous on (1, 3)
continuous on [1, 3)
y
continuous on [1, 3]
continuous on (1, 3]
y
y
y
3
3
3
3
2
2
2
2
1
1
1
1
1
2
x
3
1
2
3
x
1
2
3
x
1
2
3
x
Types of Discontinuities
When a function is not continuous at a point x = c, we say that it is discontinuous at
x = c. In terms of limits this means that the limit lim f (x) is not equal to the value f (c). The
x→c
three most basic types of discontinuities that a function can have are illustrated as follows:
removable discontinuity
jump discontinuity
y
infinite discontinuity
y
y
4
4
4
3
3
3
2
2
2
1
1
1
1
2
3
4
x
1
2
3
4
x
1
2
3
4
x
Intuitively, we say that a discontinuity is removable if we could remove it just by changing
one function value. At a jump discontinuity, the function jumps from one value to another,
and at an infinite discontinuity the function has a vertical asymptote.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
Continuity and Its Consequences
111
These types of discontinuities can be described precisely in terms of limits as follows:
DEFINITION 1.15
Removable, Jump, and Infinite Discontinuities
Suppose f is discontinuous at x = c. We say that x = c is a
(a) removable discontinuity if lim f (x) exists but is not equal to f (c);
x→c
(b) jump discontinuity if lim− f (x) and lim+ f (x) both exist but are not equal;
x→c
x→c
(c) infinite discontinuity if one or both of lim− f (x) and lim+ f (x) is infinite.
x→c
x→c
For example, in the first figure shown, the limit lim f (x) = 1 exists but is not equal to
x→2
f (2) = 2, and therefore f (x) has a removable discontinuity at x = 2. The function g(x) in
the second figure has left and right limits lim− g(x) = 1 and lim+ g(x) = 2, respectively; the
x→2
x→2
limits from both sides exist, but they are not equal to each other, and therefore h(x) has a
jump discontinuity at x = 2. Finally, the function h(x) in the third graph has an infinite limit
from both the left and the right at x = 2, and therefore has an infinite discontinuity at that
point.
Continuity of Very Basic Functions
We say that a function is continuous on its domain if it is continuous on every interval
on which it is defined. The following theorem proves that, unsurprisingly, our simplest
examples of functions are continuous on their domains:
THEOREM 1.16
Continuity of Simple Functions
(a) Constant, identity, and linear functions are continuous everywhere. In terms of
limits, for every k, c, m, and b in R we have
lim k = k,
x→c
lim x = c,
and
x→c
lim(mx + b) = mc + b.
x→c
(b) Power functions are continuous on their domains. In terms of limits, if A is real
and k is rational, then for all values x = c at which x k is defined we have
lim Ax k = Ac k .
x→c
This is a powerful theorem, because it tells us that we can calculate limits of certain simple
functions at domain points just by evaluating the functions at those points. For example,
1
f (x) =
= x−1 is a power function, so by the preceding theorem, it is continuous on
x
1
x→c x
its domain (−∞, 0) ∪ (0, ∞). This means that at any point c = 0 we can calculate lim
by simply calculating the value f (c) =
calculate
1
lim ;
x→0 x
1
.
c
However, the theorem does not tell us how to
we will discuss such limits in a later section.
Technical point: In part (b) of the theorem the limit may sometimes be only one-sided.
For example, f (x) = x 1/2 is defined at x = 0 and to the right of x = 0, but not for x < 0.
Therefore the corresponding limit statement is one-sided: lim+ x 1/2 = 0.
x→0
Although it seems graphically obvious that the simple types of functions described in
Theorem 1.16 are continuous everywhere they are defined, to actually prove continuity
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
112
Chapter 1
November 21, 2012
Limits
we need to appeal to the definition of limit. We will prove part (a) here and discuss the proof of part (b) after we learn about limit rules in the next section. You
1
1
will prove that f (x) = x k is continuous on its domain for k = 2, 3, −1, −2, , and − in
2
2
Exercises 89–92.
Proof. (This proof requires material covered in optional Section 1.3.)
The limit lim k = k makes intuitive sense because as x approaches c, the number k should simply
x→c
remain k; there is no x involved. To prove this limit statement we must show that for all > 0, there
is some δ > 0 such that if x ∈ (c − δ, c) ∪ (c, c + δ), then k ∈ (k − , k + ). But k is always in the
interval (k − , k + ), so the implication is trivially true for all values of and δ. This is illustrated
in the leftmost figure that follows.
If f (x) = k, choose any δ
If f (x) = x, choose δ = y
y
k
If f (x) = mx + b, choose δ =
y
f(c) c
k
c
cδ
c
cδ
x
rise
f(c)
rise
c
k
|m|
f(c) run
run
cδ c cδ
x
cδ c cδ
x
To prove that lim x = c we must show that for all > 0, there exists δ > 0 such that if
x→c
x ∈ (c − δ, c) ∪ (c, c + δ), then x ∈ (L − , L + ). This clearly holds if we choose δ to be equal to ,
as shown in the middle figure.
To prove that lim(mx + b) = (mc + b) we will use the definition of limit in terms of absolute
x→c
value inequalities from Definition 1.10. In the case when m = 0, the linear function f (x) = mx + b
is the constant function f (x) = b; we have already proved that case. If m = 0, then given > 0,
. Then for all x satisfying 0 < |x − c| < δ we also have
choose δ =
|m|
|(mx + b) − (mc + b)| = |mx − mc| = |m||x − c| < |m|δ = |m|
= .
|m|
rise
, or the slope
It makes intuitive sense that δ should be , since the ratio is equal to the ratio
m
δ
run
of f (x) = mx + b, as shown in the rightmost figure.
Extreme and Intermediate Values of Continuous Functions
In this section we examine two important consequences of continuity. First, a continuous
function on a closed interval must be bounded and attain its upper and lower bounds.
Second, if f is continuous between two values x = a and x = b, then the corresponding
values of f (x) go through every possible intermediate value between the y-values f (a) and
f (b). Both of these consequences are intuitively obvious if we think of continuous functions
as having “unbroken” graphs.
For example, consider the function f in the first figure that follows. This function is
continuous on [a, b]. In the second graph we see that f attains its maximum on that interval
at x = M and its minimum at x = m. In the third graph we see that the function attains
every intermediate value K between f (a) and f (b).
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
f is continuous on [a, b]
Height of y = K at x = c
Maximum at M, minimum at m
y
113
Continuity and Its Consequences
y
y
f(M)
f(a)
K
f(m)
a
f(b)
x
b
a
M
m
x
a
c
b
x
In general, attaining extreme values and passing through all intermediate values are
properties that hold for every function that is continuous on a closed interval [a, b]:
THEOREM 1.17
The Extreme Value Theorem
If f is continuous on a closed interval [a, b], then there exist values M and m in the interval
[a, b] such that f (M) is the maximum value of f (x) on [a, b] and f (m) is the minimum value
of f (x) on [a, b].
THEOREM 1.18
The Intermediate Value Theorem
If f is continuous on a closed interval [a, b], then for any K strictly between f (a) and f (b),
there exists at least one c ∈ (a, b) such that f (c) = K.
These two important consequences of continuity may seem obvious, but in fact they rely
on a subtle mathematical property of the real numbers called the Least Upper Bound
Axiom. Properly explaining the proofs of these theorems is outside of the scope of this
book.
In the Extreme and Intermediate Value Theorems, the hypothesis that f be continuous
on a closed interval [a, b] is essential. If f either fails to be continuous on the interior of
the interval or fails to be continuous at a closed endpoint, then the conclusions of these
theorems do not necessarily hold. For example, each of the following three functions fails
to be continuous on [a, b] and also fails to satisfy at least one of the conclusions of the two
theorems.
No c with f (c) = K
No minimum value on [a, b]
No maximum value on [a, b]
y
y
y
f(a)
f(a)
f(a)
K
f(b)
f(b)
a
b
x
f(b)
a
b
x
a
CONFIRMING PAGES
c
b
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
114
Chapter 1
November 21, 2012
Limits
An extremely useful special case of the Intermediate Value Theorem is the case when
we consider the intermediate value K = 0. In this case, the Intermediate Value Theorem
says that if f is continuous on [a, b] and f (a) and f (b) have opposite signs, then there exists
at least one c ∈ (a, b) where f (c) = 0. We use an equivalent variant of this special case every
time we solve an inequality by checking signs between roots and discontinuities:
THEOREM 1.19
A Function Can Change Sign Only at Roots and Discontinuities
A function f can change sign (from positive to negative or vice versa) at a point x = c
only if f (x) is zero, undefined, or discontinuous at x = c.
The graph that follows at the left shows a function f that is continuous on [a, b], and changes
sign only at its roots c 1 , c 2 , and c 3 . The graph at the right is discontinuous somewhere in
[a, b] and therefore can change sign as we move from left to right without ever crossing the
y-axis.
f (x) can change sign at roots
f (x) can change sign at a discontinuity
y
y
f(a)
f(a)
a c1
c2
c3 b
f(b)
x
a
b
x
f(b)
Examples and Explorations
EXAMPLE 1
Limits and continuity of piecewise-defined functions
Describe the continuity or discontinuity of each piecewise-defined function that follows
at x = 1 by using graphs to determine the left, right, and two-sided limits at x = 1. Then
describe the intervals on which each function is continuous.
x + 1, if x < 1
4 − x 2 , if x ≤ 1
(a) f (x) =
(b)
g(x)
=
2
3 − x , if x ≥ 1
x − 1, if x > 1
SOLUTION
(a) The graph of f looks like y = x + 1 to the left of x = 1 and like y = 3 − x 2 to the right
of x = 1, as shown next at the left. From the graph we see that lim− f (x) = 2 and
x→1
lim+ f (x) = 2, and thus lim f (x) = 2. We also have f (1) = 3 − 12 = 2. Since the limit
x→1
x→1
of f (x) as x → 1 is equal to the value of f (x) at x = 1, we can conclude that f is continuous at x = 1. In fact, according to the graph, the function f is continuous on all of
(−∞, ∞).
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
Left and right limits both equal f (1)
Only the left limit equals g(1)
y
y
3
4
2
3
1
2
−3 2 1
1
115
Continuity and Its Consequences
1
2
3
x
1
2
3 2 1
1
3
2
1
2
x
3
(b) The graph of g(x) looks like 4 − x 2 to the left of x = 1 and like x − 1 to the right of
x = 1, with value g(1) = 4 − 12 = 3, as shown previously at the right. From the graph
we see that lim− g(x) = 3 while lim+ g(x) = 0. The left and right limits both exist, but
x→1
x→1
they are not equal; thus g(x) has a jump discontinuity at x = 1. Since lim− g(x) = 3 =
x→1
g(1) but lim+ g(x) = 0 = g(1), we can also say that g(x) is left continuous, but not
x→1
right continuous, at x = 1. According to the graph, the function g(x) is continuous on
(−∞, 1] and on (1, ∞).
EXAMPLE 2
Continuity of a function that is defined separately for rationals and irrationals
Determine graphically whether or not the rather exotic functions that follow are continuous
at x = 0. Although these types of functions are not going to be a major focus in this course,
this example helps get at the root of what continuity really means. You might be surprised
by the solution!
1, if x is rational
1, if x is rational
(b) g(x) =
(a) f (x) =
x + 1, if x is irrational
−1, if x is irrational
SOLUTION
In the graphs of f and g that follow, the lighter dotted line represents the values of the
function at rational-number inputs and the darker dotted line represents the values at
irrational-number inputs. Note that the graphs of both f and g pass the vertical line test,
since every input x is either rational or irrational and never both.
Overall limit does not exist as x → 0
Limit as x → 0 approaches 1 in both cases
y
2
y
2
3
1
2
1
1
1
2
2
x
1
2
1
1
2
x
1
(a) We must consider the limit of f as x → 0 separately for rational and irrational values
of x. From the graph at the left, we see that for rational values of x we have lim f (x) = 1
x→0
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
116
Chapter 1
November 21, 2012
Limits
while for irrational values of x we have lim f (x) = −1. Note that every punctured interx→0
val (−δ, 0) ∪ (0, δ) around x = 0 contains both rational and irrational numbers. Since
the limit of f (x) as x → 0 is different depending on whether we choose rational or
irrational values of x, the overall limit does not exist. Therefore the function f is not
continuous at x = 0.
(b) On the other hand, looking at the graph of g at the right, we see that for rational values
of x we have lim g(x) = 1 and for irrational values of x we also have lim g(x) = 1.
x→0
x→0
Therefore lim g(x) exists and is equal to 1. Since g(0) is also equal to 1, the function g
x→0
is in fact continuous at x = 0, as strange as that may seem.
EXAMPLE 3
Calculating limits of very basic functions
Use continuity to calculate each of the limits that follow, if possible. If we do not yet have
enough information to calculate a limit, explain why not.
(a) lim 2
x→3
1
x→3 x
(b) lim
1
x→0 x
(c) lim
(d) lim
x→4
√
x−1
SOLUTION
By Theorem 1.16 we know that the constant function f (x) = 2 and the power function
1
g(x) = are continuous on their domains. Therefore for parts (a) and (b) we can calculate
x
1
x→3 x
the limits just by evaluating at x = 3: lim 2 = 2 and lim
x→3
1
3
= .
We do not yet know how to calculate the remaining two limits algebraically. In part (c),
1
the point x = 0 is not in the domain of g(x) = , so we cannot apply Theorem 1.16. In
x
√
part (d), the function h(x) = x − 1 is not a constant, identity, linear, or power function, and thus at this point we cannot conclude anything about its continuity or its
limits.
EXAMPLE 4
A real-world illustration of the Extreme and Intermediate Value Theorems
Consider the function w(t) that describes a particular person’s weight at t years of age
between the ages of 18 and 45. Why does it make sense that this function is continuous
on [18, 45]? What do the Extreme Value Theorem and the Intermediate Value Theorem say
about w(t)?
SOLUTION
The weight function w(t) should be continuous on [18, 45] because a person’s weight
changes continuously over time and cannot jump from one value to another. (We are
assuming typical circumstances, so that a person does not get a serious haircut, lose a limb,
or somehow otherwise get their weight to change drastically in an instant.)
The Extreme Value Theorem tells us that there is some time M ∈ [18, 45] at which the
person’s weight was greatest and some time m ∈ [18, 45] at which that person weighed the
least. In other words, at some time between 18 and 45 years of age, the person must have
had a maximum weight and a minimum weight.
The Intermediate Value Theorem tells us that for every weight K between w(18)
and w(45), there is some time c ∈ (18, 45) for which w(c) = K. For example, if the person weighed w(18) = 130 pounds at age 18 and w(45) = 163 pounds at age 45, then
there must be some age between 18 and 45 at which the person weighed, say, exactly
144 pounds.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
EXAMPLE 5
Continuity and Its Consequences
117
Applying the Intermediate Value Theorem to a Continuous Function
The function f (x) = x 3 − 3x + 1 is continuous everywhere. (We will see this later in
Section 1.5.) Use the Intermediate Value Theorem to conclude that there is some point c
for which f (c) = 2. Then use a graph of f to approximate at least one such value of c.
SOLUTION
To show that there is some c with f (c) = 2 we need to find values a and b such that K = 2
is between f (a) and f (b), and apply the Intermediate Value Theorem. By trial and error we
can find such values a and b, by testing different values of f (x) until we find one that is less
than and one that is greater than 2. For example,
f (0) = 03 − 3(0) + 1 = 1 < 2,
f (2) = 23 − 3(2) + 1 = 3 > 2.
Since f is continuous on [0, 2] and f (0) < 2 < f (2), by the Intermediate Value Theorem
there is some value c ∈ (0, 2) for which f (c) = 2. Note that the Intermediate Value Theorem doesn’t tell us where c is, only that such a c exists somewhere in the interval
(0, 2).
We can approximate some values of c for which f (c) = 2 by approximating the values
of x for which the graph of f (x) = x 3 − 3x + 1 intersects the line y = 2:
3.5
2.5
2.5
1.5
From this graph we can conclude that f (c) = 2 at c ≈ −1.5, c ≈ −0.4, and c ≈ 1.9. To
get better approximations we could trace along the graph on a calculator or other graphing
utility.
EXAMPLE 6
Determining sign information between zeroes and discontinuities
Determine the intervals on which the functions that follow are positive or negative. You
may assume that the function in part (a) is continuous everywhere and the function in
part (b) is continuous on each piece.
(a) f (x) = 3x 3 + 3x 2 − 6x
(b) g(x) =
⎧
⎨
x + 1, if x < −2
(x + 1)2 , if − 2 ≤ x ≤ 1
⎩
2 − x, if x > 1
SOLUTION
(a) The roots of f (x) = 3x 3 + 3x 2 − 6x = 3x(x 2 + x − 2) = 3x(x − 1)(x + 2) are x = 0,
x = 1, and x = −2. By Theorem 1.19, in each of the intervals between these points
the function f is either always positive or always negative. We need to test the sign of
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
118
Chapter 1
November 21, 2012
Limits
f (x) at one point in each interval. For example, f (−3) = −36 < 0, f (−1) = 6 > 0,
f (0.5) = −1.875 < 0, and f (2) = 24 > 0. Reading from the resulting sign chart shown
at the left, we can see that f (x) = 3x 3 + 3x 2 − 6x is negative on (−∞, −2) ∪ (0, 1) and
positive on (−2, 0) ∪ (1, ∞):
2
0
1
f
dc 2
dc 1
1
2
g
(b) The piecewise-defined function g can be discontinuous only at its break points x = −2
and x = 1. Furthermore, its first component x+1 is never zero on (−∞, −2), its second
component (x + 1)2 is zero only at x = −1, and its third component 2 − x is zero
when x = 2. By Theorem 1.19 the function g can change sign only at the roots and
discontinuities at −2, −1, 1, and 2. All that now remains is to check the sign of g(x)
one time between each of these points; the results are recorded on the preceding sign
chart at the right. We marked the discontinuous points with “dc” to distinguish them
from the zeros. Reading from this sign chart and keeping careful track of the sign of
g(x) at the break points, we see that g(x) is negative on (−∞, −2) ∪ (2, ∞) and positive
on [−2, −1) ∪ (−1, 2].
CHECKING
THE ANSWER
We can graph f and g with a calculator or other graphing utility to verify that the sign charts
we found are reasonable. Notice that the intervals where f (x) or g(x) are positive are the
intervals on which their graphs are above the x-axis.
f (x) is above the x-axis on (−2, 0) ∪ (1, ∞)
g(x) is above the x-axis on [−2, −1) ∪ (−1, 2)
4
25
3
2
3
22
15
TEST YOUR
? UNDERSTANDING
3
Use limits to give definitions of each of the following: continuity at a point, continuity
on an interval, left and right continuity.
Use limits to give definitions of each of the following: removable discontinuity, jump
discontinuity, infinite discontinuity.
In the reading is a graph of a function with a removable discontinuity. We could make
this graph continuous if we could change just one function value. What value?
Why would it be difficult to prove at this point that all power functions, no matter what
kind of power, are continuous on their domains?
The conclusion of the Intermediate Value Theorem tells us of the existence of a point
c ∈ (a, b) with f (c) = K. Why do we need only the open interval (a, b), and not the
closed interval [a, b], in this conclusion?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
Continuity and Its Consequences
119
EXERCISES 1.4
Thinking Back
Finding roots of piecewise-defined functions: For each function f
that follows, find all values x = c for which f (c) = 0. Check
your answers by sketching a graph of f .
4 − x 2 , if x < 0
f (x) =
x + 1, if x ≥ 0
x + 1, if x < 0
f (x) =
4 − x 2 , if x ≥ 0
2x − 1, if x ≤ 1
f (x) =
2x 2 + x − 3, if x > 1
Logical existence statements: Determine whether each of the
statements that follow are true or false. Justify your answers.
If x is an integer, then there exists some positive
integer y such that |y| = x.
If x is a positive integer, then there exists some
negative integer y such that |y| = x.
If x ∈ [−2, 2], then there exists some y ∈ (0, 4) such
that y = x 2 .
If x ∈ [0, 100], then there exists some y ∈ [−10, 10]
such that x = y2 .
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If f is both left and right continuous at
x = c, then f is continuous at x = c.
(b) True or False: If f is continuous on the open interval
(0, 5), then f is continuous at every point in (0, 5).
(c) True or False: If f is continuous on the closed interval
[0, 5], then f is continuous at every point in [0, 5].
(d) True or False: If f is continuous on the interval (2, 4),
then f must have a maximum value and a minimum
value on (2, 4).
(e) True or False: If f (3) = −5 and f (9) = −2, then there
must be a value c at which f (c) = −3.
(f) True or False: If f is continuous everywhere, and if
f (−2) = 3 and f (1) = 2, then f (x) must have a root
somewhere in (−2, 1).
(g) True or False: If f is continuous everywhere, and if
f (0) = −2 and f (4) = 3, then f (x) must have a root
somewhere in (0, 4).
(h) True or False: If f (0) = f (6) = 0 and f (2) > 0, then f (x)
is positive on the entire interval (0, 6).
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function with f (4) = 2 that has a
removable discontinuity at x = 4.
(b) The graph of a function that is continuous on its
domain but not continuous at x = 0.
(c) The graph of a function that is continuous on (0, 2]
and (2, 3) but not on (0, 3).
3. If f is a continuous function, what can you say about
lim f (x)?
x→1
4. Explain what it means for a function f to be continuous
at a point x = c, with a sentence that includes the words
“approaches” and “value.”
5. In our proof that constant functions are continuous, we
used the fact that given any > 0, a choice of any δ > 0
will work in the formal definition of limit. Use a graph to
explain why this makes intuitive sense. (This exercise depends on Section 1.3.)
6. In our proof that linear functions are continuous, we used
will
the fact that given any > 0, the choice of δ =
|m|
work in the formal definition of limit. Use a graph to explain why this makes intuitive sense. (This exercise depends
on Section 1.3.)
7. Given the following function f , define f (1) so that f is continuous at x = 1, if possible:
f (x) =
x 2 − 2x + 1
.
x 2 − 6x + 5
8. Given the following function f , define f (1) so that f is continuous at x = 1, if possible:
3x − 1, if x < 1
f (x) =
x 2 + 1, if x > 1.
Each function in Exercises 9–12 is discontinuous at some value
x = c. Describe the type of discontinuity and any one-sided
continuity at x = c, and sketch a possible graph of f .
9.
10.
11.
12.
lim f (x) = 2, lim + f (x) = 2, f (−1) = 1.
x→−1−
x→−1
lim f (x) = 2, lim+ f (x) = 1, f (2) = 1.
x→2−
x→2
lim− f (x) = −1, lim+ f (x) = 1, f (0) = 0.
x→0
x→0
lim− f (x) = −∞, lim+ f (x) = ∞, f (2) = 3.
x→2
x→2
13. State what it means for a function f to be continuous at
a point x = c, in terms of the delta–epsilon definition of
limit. (This exercise depends on Section 1.3.)
14. State what it means for a function f to be left continuous
at a point x = c, in terms of the delta–epsilon definition
of limit. (This exercise depends on Section 1.3.)
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
120
Chapter 1
November 21, 2012
Limits
15. State what it means for a function f to be right continuous
at a point x = c, in terms of the delta–epsilon definition
of limit. (This exercise depends on Section 1.3.)
16. Sketch a labeled graph of a function that satisfies the
hypothesis of the Extreme Value Theorem, and illustrate
on your graph that the conclusion of the Extreme Value
Theorem follows.
19. Sketch a labeled graph of a function that fails to satisfy the
hypothesis of the Extreme Value Theorem, and illustrate
on your graph that the conclusion of the Extreme Value
Theorem does not necessarily hold.
20. Explain why the Intermediate Value Theorem allows us to
say that a function can change sign only at discontinuities
and zeroes.
17. Sketch a labeled graph of a function that satisfies the
hypothesis of the Intermediate Value Theorem, and
illustrate on your graph that the conclusion of the Intermediate Value Theorem follows.
18. Sketch a labeled graph of a function that fails to satisfy
the hypothesis of the Intermediate Value Theorem, and
illustrate on your graph that the conclusion of the Intermediate Value Theorem does not necessarily hold.
For each of the following sign charts, sketch the graph of a
function f that has the indicated signs, zeros, and discontinuities:
21.
22.
0
4
1
3
dc dc 2
0
2
f
f
Skills
For each function f graphed in Exercises 23–26, describe the
intervals on which f is continuous. For each discontinuity
of f , describe the type of discontinuity and any one-sided
continuity. Justify your answers about discontinuities with
limit statements.
23. y
y
24.
35.
2
3
1
2
3 2 1
1
1
1
2
3
x
2
1
2
4
3
x
y
25.
3
y
26.
3
3
2
2
1
1
3 2 1
1
2
3
x
3 2 1
1
1
2
2
3
For each limit in Exercises 33–38, either use continuity to calculate the limit or explain why Theorem 1.16 does not apply.
33.
3
4
31. f has a removable discontinuity at x = −2 and is right
continuous at x = −2, and f (−2) = 0.
32. f is continuous on [0, 2) but not on [0, 2].
1
2
3
x
Sketch the graph of a function f described in Exercises 27–32,
if possible. If it is not possible, explain why not.
27. f is left continuous at x = 1 and right continuous at x = 1,
but is not continuous at x = 1, and f (1) = −2.
28. f is left continuous at x = 2 but not continuous at x = 2,
and f (2) = 3.
29. f has a jump discontinuity at x = −1 and is left continuous at x = −1, and f (−1) = 2.
30. f has an infinite discontinuity at 0 but is right continuous
at 0, and f (0) = 1.
lim 6
34.
lim (3x − 2)
36. lim x 4
x→3
√
38. lim x
x→−1
x→−5
37. lim x−3
x→0
lim x
x→−1
x→−5
In Exercises 39–44, use Theorem 1.16 and left and right limits to determine whether each function f is continuous at its
break point(s). For each discontinuity of f , describe the type
of discontinuity and any one-sided discontinuity.
x − 3, if x < 3
39. f (x) =
−(x − 3), if x ≥ 3
x − 3, if x < 0
40. f (x) =
−(x − 3), if x ≥ 0
⎧
x 2 , if x < 2
⎨
4, if x = 2
41. f (x) =
⎩
2x + 1, if x > 2
⎧√
⎨ −x, if x < 0
2, if x = 0
42. f (x) =
⎩ √
x, if x > 0
⎧
⎨ x + 1, if x < 1
43. f (x) = 3x − 1, if 1 ≤ x < 2
⎩
x + 2, if x ≥ 2
⎧
⎨ x 3 , if x ≤ 0
44. f (x) = 1 − x, if 0 < x < 3
⎩
x − 5, if x ≥ 3
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.4
Continuity and Its Consequences
Use graphs to determine if each function f in Exercises 45–48
is continuous at the given point x = c.
2 − x, if x rational
c=2
45. f (x) =
x 2 , if x irrational,
2
x − 3, if x rational
46. f (x) =
c=0
3x + 1, if x irrational,
2 − x, if x rational
c=1
47. f (x) =
x 2 , if x irrational,
2
x − 3, if x rational
48. f (x) =
c=4
3x + 1, if x irrational,
59. f (x) = x 3 − 3x 2 − 2, [a, b] = [2, 4], K = −4
Use the Extreme Value Theorem to show that each function f
in Exercises 49–54 has both a maximum and a minimum value
on [a, b]. Then use a graphing utility to approximate values M
and m in [a, b] at which f has a maximum and a minimum,
respectively. You may assume that these functions are continuous everywhere.
63. f (x) = sin x, K =
49. f (x) = x 4 − 3x 2 − 2, [a, b] = [−2, 2]
121
60. f (x) = 2 + x + x 3 , [a, b] = [−1, 2], K = 3
Use the Intermediate Value Theorem to show that for each
function f and value K in Exercises 61–66, there must be some
c ∈ R for which f (c) = K. You will have to select an appropriate interval [a, b] to work with. Then find or approximate
one such value of c. You may assume that these functions are
continuous everywhere.
61. f (x) = x 3 + 2, K = −15
62. f (x) = −2x 2 + 4, K = 0
1
2
√
3
64. f (x) = sin x, K =
2
65. f (x) = |3x + 1|, K = 1
66. f (x) = |2 − 3x|, K = 2
52. f (x) = 3 − 2x 2 + x 3 , [a, b] = [−1, 2]
Find the intervals on which each function in Exercises 67–74
is positive or negative. Make clear how your work uses the
Intermediate Value Theorem and continuity. You may assume
that polynomials and their quotients are continuous on the
intervals on which they are defined.
53. f (x) = 3 − 2x 2 + x 3 , [a, b] = [0, 2]
67. f (x) = 2 + 5x + 2x 2
50. f (x) = x 4 − 3x 2 − 2, [a, b] = [0, 2]
51. f (x) = x 4 − 3x 2 − 2, [a, b] = [−1, 1]
54. f (x) = 3 − 2x + x , [a, b] = [−1, 1]
2
3
Use the Intermediate Value Theorem to show that for each
function f , interval [a, b], and value K in Exercises 55–
60, there is some c ∈ (a, b) for which f (c) = K. Then use a
graphing utility to approximate all such values c. You may
assume that these functions are continuous everywhere.
71.
72.
73.
57. f (x) = x 3 − 3x 2 − 2, [a, b] = [−2, 4], K = −4
58. f (x) = x 3 − 3x 2 − 2, [a, b] = [0, 2], K = −4
(x + 4)(x − 1)
x2 − 4
70. f (x) =
2x + 3
x2 − 1
x − 4, if x ≤ 1
f (x) =
x 2 − 4, if x > 1
3x + 1, if x < 0
f (x) =
x, if x ≥ 0
x 3 , if x ≤ 2
f (x) =
4x − x 3 , if x > 2
x 2 − 9, if x ≤ −2
f (x) =
x 2 + x − 2, if x > −2
69. f (x) =
55. f (x) = 5 − x 4 , [a, b] = [0, 2], K = 0
56. f (x) = 5 − x 4 , [a, b] = [−2, −1], K = 0
68. f (x) = x 3 − 2x 2 − 3x
74.
Applications
Explain in practical terms what the Extreme Value Theorem says about each continuous function defined in Exercises
75–77. Then explain in practical terms what the Intermediate
Value Theorem says in each situation.
75. Alina hasn’t cut her hair for six years. Six years ago her hair
was just 2 inches long. Now her hair is 42 inches long. Let
H(t) be the function that describes the length, in inches,
of Alina’s hair t years after she stopped cutting it.
76. Linda collects rain in a bucket outside her back door.
Since the first day of April she has been keeping track
of how the amount of water in the bucket changes as it
fills with rain and evaporates. On April 1 the bucket was
empty, and today it contains 4 inches of water. Let w(t)
be the height, in inches, of rainwater in the bucket t days
after the first day of April.
77. The number of gallons of gas in Phil’s new station wagon t
days after he bought it is given by the function g(t). When
he purchased the station wagon one year ago, the tank
had 19 gallons of gas in it. Today he ran out of gas.
78. Lars was 20 inches tall when he was born, and six foot
one when he died at age 83. Use the Intermediate Value
Theorem to show that there must have been some point
in Lars’s life at which his height in inches was equal to his
age in years. (Hint: Think about when the difference between
his height and age is zero.)
79. As a vacuum cleaner salesman, Alex earns a salary of
$8,500 a year, whether he sells any vacuum cleaners or
not. In addition, for every 30 vacuum cleaners he sells, he
earns a $1,500 commission.
(a) Construct a piecewise-defined function M(v) that describes the amount of money M that Alex will make in
a year if he sells v vacuum cleaners over the course of
the year. Assume he sells between 0 and 90 vacuum
cleaners in a year.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
122
Chapter 1
November 21, 2012
Limits
(b) Check that your function makes sense by using
it to calculate M(0), M(30), M(59), M(61), and
M(90). Then sketch a graph of M(v) on the interval
0 ≤ v ≤ 90.
(c) The piecewise-defined function M(v) is not continuous. List all the values at which M(v) fails to be
continuous, and use the definition of continuity to
support your answers.
80. One immediate application of the Intermediate Value
Theorem is the method of finding roots called the
Bisection Method. In this problem you will develop this
method and then use it to approximate the square root
of 2.
(a) Suppose f is continuous on R and that a and b are
some real numbers for which f (a) is negative and f (b)
is positive. Explain why the Intermediate Value Theorem guarantees that there must be some point in (a, b)
where f (x) has a root.
(b) Consider the function f (x) = x 2 − 2. Show that f (0)
is negative and f (2) is positive. What conclusion can
we draw from the Intermediate Value Theorem?
(c) We can bisect the interval (0, 2) by finding the midpoint of the interval, which in this case is x = 1.
Is f (1) positive or negative? Does the Intermediate Value Theorem say anything about f (x) = x 2 − 2
on the interval (0, 1)? What about on the interval
(1, 2)?
(d) Your answer to part (b) tells you that f (x) = x 2 − 2
must have a root somewhere in the interval (0, 2)
of length 2. Your answer to part (c) tells you that
f (x) = x 2 − 2 must have a root in a shorter interval
of length 1. Now repeat! Bisect the interval of length
1
1 to find an interval of length on which f (x) must
2
have a root.
(e) Describe why this Bisection Method will in general
give better and better approximations for finding a
root of a given function. In this particular example,
with f (x) = x 2 − 2, why does√the Bisection Method
give us an approximation for 2?
Proofs
81. Write a delta–epsilon proof that shows that the function
f (x) = 3x − 5 is continuous at x = 2. (This exercise depends
on Section 1.3.)
82. Write a delta–epsilon proof that shows that the function
f (x) = 2x + 1 is continuous at x = 5. (This exercise depends
on Section 1.3.)
83. Write a delta–epsilon proof that shows that the function f (x) = |x| is continuous. You may find the following
inequality useful: For any real numbers a and b,
||a| − |b|| ≤ |a − b|. (This exercise depends on Section 1.3.)
84. Use what you know about one-sided limits to prove that
a function f is continuous at a point x = c if and only if it
is both left and right continuous at x = c.
For each function f in Exercises 85–86, use the delta–epsilon
definition of continuity to argue that f is or is not continuous
at the indicated point x = c.
2 − x, if x rational
c=2
85. f (x) =
x 2 , if x irrational,
Thinking Forward
Interesting trigonometric limits: For each of the functions that
follow, use a calculator or other graphing utility to examine
the graph of f near x = 0. Does it appear that f is continuous
at x = 0? Make sure your calculator is set to radian mode.
f (x) =
1
sin(x), if x = 0
x
1, if x = 0
86. f (x) =
2 − x, if x rational
c=1
x 2 , if x irrational,
87. Use the Intermediate Value Theorem to prove that every
cubic function f (x) = Ax 3 + Bx 2 + Cx + D has at least one
real root. You will have to first argue that you can find
real numbers a and b so that f (a) is negative and f (b) is
positive.
For each power function f in Exercises 89–93, write a delta–
epsilon proof which proves that f is continuous on its domain.
In each case you will need to assume that δ is less than or equal
to 1. (These exercises depend on Section 1.3.)
88. f (x) = x−1
89. f (x) = x 2
90. f (x) = x 3
91. f (x) = x −2
92. f (x) = x−1/2
93. f (x) = x 1/2
⎧ 1
⎨
sin
, if x = 0
f (x) =
x
⎩
0, if x = 0
⎧
1
⎨
x sin
, if x = 0
f (x) =
x
⎩
0, if x = 0
⎧
1
⎨ 2
x sin
, if x = 0
f (x) =
x
⎩
0, if x = 0
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
1.5
Limit Rules and Calculating Basic Limits
123
LIMIT RULES AND CALCULATING BASIC LIMITS
Rules for calculating limits of arithmetic combinations and compositions of functions
Continuity of algebraic and transcendental functions at domain points
The Cancellation Theorem and the Squeeze Theorem for calculating limits
Limits of Combinations of Functions
Although we now understand in depth what limit statements mean, at this point we do not
have many tools for calculating limits. We can calculate limits of continuous functions at
domain points by evaluation, and we know that very simple functions, such as constant and
linear functions, are continuous. What about more complicated functions? For example, we
already know from the continuity of power functions that
lim x 2 = 22 = 4 and lim x 3 = 23 = 8.
x→2
x→2
Can we use these results to say something about lim (x 2 +x 3 ) of the sum of these functions?
x→2
The key theorem that follows will help us answer this question; it says that limits behave
well with respect to all of the arithmetic operations and even with respect to composition.
THEOREM 1.20
Rules for Calculating Limits of Combinations
If lim f (x) and lim g(x) exist, then the following rules hold for their combinations:
x→c
x→c
Constant Multiple Rule: lim kf (x) = k lim f (x), for any real number k.
x→c
x→c
Sum Rule: lim( f (x) + g(x)) = lim f (x) + lim g(x)
x→c
x→c
x→c
Difference Rule: lim( f (x) − g(x)) = lim f (x) − lim g(x)
x→c
x→c
x→c
Product Rule: lim( f (x)g(x)) = (lim f (x))(lim g(x))
x→c
x→c
x→c
lim f (x)
f (x)
x→c
=
, if lim g(x) = 0
x→c g(x)
x→c
lim g(x)
Quotient Rule: lim
x→c
Composition Rule: lim f ( g(x)) = f (lim g(x)), if f is continuous at lim g(x)
x→c
x→c
x→c
This theorem is a powerful tool for calculating limits, since it tells us how to find limits
of compound functions in terms of the limits of their components. For example, we can
calculate the limit of the sum x 2 + x 3 as x → 2 by taking the sum of the limits of x 2 and
x 3 as x → 2:
lim (x 2 + x 3 ) = lim x 2 + lim x 3 = 4 + 8 = 12.
x→2
x→2
x→2
We will postpone the proofs of the limit rules in Theorem 1.20 until the end of this
section so that we can first explore their consequences and practical uses. For example,
an immediate consequence of Theorem 1.20 is that constant multiples, sums, differences,
products, quotients, and compositions of continuous functions are continuous:
THEOREM 1.21
Combinations of Continuous Functions Are Continuous
If f and g are continuous at x = c and k is any constant, then the functions kf , f + g,
f − g, and fg are also continuous at x = c.
Moreover, if g(c) = 0, then
f
g
is continuous at x = c, and if f is also continuous at g(c),
then f ◦ g is continuous at x = c.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
124
Chapter 1
November 21, 2012
Limits
For example, since f (x) = x 2 and g(x) = x 3 are continuous at x = 2, Theorem 1.21 tells
us that ( f + g)(x) = x 2 + x 3 must be also be continuous at x = 2. This makes sense given
Theorem 1.20 because
lim ( f + g)(x) = lim (x 2 + x 3 ) = lim x 2 + lim x 3 = 22 + 23 = 4 + 8 = 12 = ( f + g)(2).
x→2
x→2
x→2
x→2
Limits of Algebraic Functions
With the limit rules we can now prove that most of the functions we will use in this book
are continuous on their domains. We will start with the algebraic functions. Recall that
a function is algebraic if it can be expressed with the use of only arithmetic operations
(+, −, ×, and ÷) and rational constant powers. Power functions, polynomial functions,
and rational functions are all examples of algebraic functions.
THEOREM 1.22
Continuity of Algebraic Functions
All algebraic functions are continuous on their domains. In particular, if x = c is in the
domain of an algebraic function f , then we can calculate lim f (x) by evaluating f (c).
x→c
With this theorem we can do lots of basic limit calculations. For example, lim x 1/2 =
4
4
2, lim (3x − 2x) = 3(1) − 2(1) = 1, and
x→1
1+x
lim
x→2 3 −
√x
the limits are only one-sided; for example, lim+
x→2
=
1+2
3−2
x→4
√
4=
= 3. For certain special cases
x − 2 = 0. Note that the theorem does
not tell us how to calculate limits at non-domain points; for example, we still do not know
1
how to calculate lim
.
x→2 x − 2
Proof. Algebraic functions are by definition built out of rational powers and arithmetic combinations of real numbers and the variable x. We already know how to handle limits of constant multiples, sums, products, quotients, and compositions by using the limit rules. We also know from
Theorem 1.16 that lim k = k and lim x = c. Therefore to show that every algebraic function is conx→c
x→c
tinuous on its domain, it suffices to show that every function of the form f (x) = x k is continuous
on its domain.
We must show that for any rational number k, if x = c is in the domain of x k , then lim x k = c k .
x→c
There are a few cases to consider. If k is a positive integer, then we just repeatedly apply the product
rule for limits so that we can use the known limit lim x = c:
x→c
lim x k = (lim x)(lim x) · · · (lim x) = (c)(c) · · · (c) = c k .
x→c
x→c
x→c
x→c
k times
k times
For negative integer powers, we apply the quotient rule for limits and the result for positive integer
powers. In this case we must require c = 0 so that c will be in the domain of x−k , and we obtain the
following limit:
lim 1
1
1
x→c
lim x−k = lim k =
= k = c−k .
x→c
x→c x
c
lim x k
x→c
Although we will not prove so here, it can be shown that lim x 1/q = c 1/q when c is in the domain
x→c
of x 1/q . Given these facts, the composition rule for limits allows us to prove that x p/q is continuous
p
at domain points for any rational power :
q
lim x
x→c
p/q
= lim
x→c
√
√
q
q p
x = q lim x p = c p = c p/q .
x→c
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
Limit Rules and Calculating Basic Limits
125
Finding Limits by Cancelling or Squeezing
The continuity of algebraic functions and the limit rules can help us calculate a great many
limits, but only at domain points. One thing that can help us at non-domain points is the
x2 − 1
. At x = 1 we
x→1 x − 1
0
form . Limits of this form
0
cancellation of common factors. For example, consider the limit lim
have x 2 − 1 = 0 and x − 1 = 0, and therefore the limit is of the
are said to be indeterminate, which means that they may or may not exist, depending on
the situation. We will examine indeterminate forms in depth in Section 1.6. In the example
we are considering, we can determine the limit by simple cancellation:
x2 − 1
(x − 1)(x + 1)
= lim
= lim (x + 1) = 2.
x−1
x→1 x − 1
x→1
x→1
lim
The cancellation of the common factor x − 1 is valid because
(x − 1)(x + 1)
x−1
and x + 1 differ
only when x = 1, and when we take the limit, we are not concerned with what happens at
the point x = 1. In general, by definition, limits as x → c never have anything to do with
what happens at x = c, which proves the following theorem:
THEOREM 1.23
The Cancellation Theorem for Limits
If lim g(x) exists, and f is a function that is equal to g for all x sufficiently close to c except
x→c
possibly at c itself, then lim f (x) = lim g(x).
x→c
x→c
Another useful tool for calculating new types of limits is the Squeeze Theorem. This
theorem says that if the output of a function f (x) is always bounded between a lower function l(x) and an upper function u(x), and if the lower and upper functions approach the
same value L as x → c, then f (x) gets squeezed between the lower and upper functions
and also approaches L as x → c.
THEOREM 1.24
The Squeeze Theorem for Limits
If l(x) ≤ f (x) ≤ u(x) for all x sufficiently close to c, but not necessarily at x = c, and if
lim l(x) and lim u(x) are both equal to L, then lim f (x) = L.
x→c
x→c
x→c
Similar results hold for limits at infinity and one-sided limits.
For example, the figure that follows shows a function f that is sandwiched between two
functions u and l as x → 0. Since u(x) and l(x) have the same limit at x → 0 and f (x) is
squeezed between them, we know that f (x) must share that same limit as x → 0.
l(x) ≤ f (x) ≤ u(x)
y
u(x)
3
2
1
4 3 2 1
1
f(x)
1
2
3
4
x
2
3
l(x)
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
126
Chapter 1
November 21, 2012
Limits
Proof. Given > 0, we can choose δ1 > 0 to get u(x) within of L and also choose δ2 to get l(x)
within of L. If δ = min(δ1 , δ2 ), then whenever x ∈ (c − δ, c) ∪ (c, c + δ), we also have
L − < l(x) ≤ f (x) ≤ u(x) < L + .
Similar arguments prove that the Squeeze Theorem holds for limits as x → ∞ and x → c+ or
x → c− .
For example, we can use the Squeeze Theorem to calculate the limits of sin θ and cos θ at
x = 0. Consider the following diagrams of portions of the unit circle, with angles measured
π
in radians and 0 < θ < :
4
0 < 1 − cos θ < θ
0 < sin θ < θ
y
y
sin θ
θ
θ0
θ0
θ
x
x
cos θ
1 cos θ
According to the leftmost figure, we clearly have 0 < sin θ < θ . Since lim+ 0 = 0 and
θ→0
lim+ θ = 0, by the Squeeze Theorem we must also have lim+ sin θ = 0. In the second
θ→0
θ→0
figure we can see that 0 < 1 − cos θ < θ. Thus again by the Squeeze Theorem we
have lim+ (1 − cos θ ) = 0. Rewriting this limit with some simple limit rules, we see that
θ→0
lim+ cos θ = 1. We can illustrate similar limits as θ → 0− by drawing pictures of small
x→0
negative values of θ in the fourth quadrant (see Exercises 22 and 23).
Defining the Number e
Before we can discuss continuity and limits of exponential functions, we must have a
proper definition for the irrational number e that we have so far been approximating as
e ≈ 2.71828. It turns out that this definition itself involves a limit.
DEFINITION 1.25
The Number e
We define e to be the number that (1 + h)1/ h approaches as h approaches 0:
e = lim (1 + h)1/ h .
h→0
It is important to note that this is a weak definition, because we have not proven that the
limit in this definition exists. That is, we have not shown that (1 + h)1/ h converges to a real
number as h → 0.
For small values of h it is easy to see that the quantity (1 + h)1/ h is close to the approximation e ≈ 2.71828 we have been using so far. For example, when h = 0.0001, we have
(1 + 0.0001)1/0.0001 ≈ 2.71815.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
Limit Rules and Calculating Basic Limits
127
Proving that the limit in Definition 1.25 does converge to a real number is beyond the scope
of this chapter. In Example 5 and Exercise 24 we will use tables of values to show that this
limit is reasonable. We will use accumulation integrals to give another definition of the
number e in Section 4.7.
As we will see in the next chapter, our definition of the number e is partially motivated
by derivatives. Specifically, the reason that e is the natural base for exponential functions
has to do with the slope of the graph of y = e x at x = 0. In the next chapter we will see
that the slope of the graph of an exponential function y = b x at x = 0 is given by the limit
bh −1
.
h
h→0
lim
THEOREM 1.26
We can use our definition of e to show that when b = e, this slope is equal to 1:
Another Characterization of the Number e
The number e satisfies the following limit statement:
eh − 1
= 1.
h
h→0
lim
Proof. For a proper proof of this limit, we need a technique that we will not cover until Section 3.6. Thus we give here only a convincing argument that uses approximations. Given that
e = lim (1 + h)1/ h as in Definition 1.25, for sufficiently small values of h we have
h→0
e ≈ (1 + h)1/ h
⇐⇒
eh ≈ 1 + h
⇐⇒
eh − 1 ≈ h
⇐⇒
eh − 1
≈1
h
e h −1
= 1.
h
h→0
Since the preceding approximations get better as h → 0, it is reasonable that lim
Continuity of Exponential and Trigonometric Functions
We can use the definition of e to prove that exponential and logarithmic functions
are continuous everywhere they are defined. This allows us to calculate limits such as
lim 3 x = 32 = 9 and lim ln x = ln 1 = 0 by simple evaluation.
x→2
THEOREM 1.27
x→1
Continuity of Exponential and Logarithmic Functions
All exponential and logarithmic functions are continuous on their domains.
Proof. The proof hinges entirely on algebra, limit rules, and the definition of e in Definition 1.25.
We will prove continuity for (a) the natural exponential function and (b) the natural logarithmic
function here, and use limit rules to extend these results to general exponential and logarithmic
functions in Exercises 94–96. The proofs are a little technical, but without them we would not be
able to compute even the simplest limits of exponential and logarithmic functions!
(a) To prove that e x is continuous on its domain R we must show that for all c ∈ R we have
lim e x = e c . If we want to use the definition of e, then we need to have a limit as h → 0, so we
x→c
define h = x − c. Then x = c + h, and as x → c, we have h → 0. This makes our limit equal to
lim e x = lim e c+h = lim e c e h = e c lim e h .
x→c
h→0
h→0
h→0
The last step follows from the constant multiple rule for limits, since e c is a constant. At this
point we would be done if we could show that lim e h = e0 = 1. In other words, the proof that
h→0
e x is continuous for all x essentially boils down to showing that it is continuous at one point,
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
128
Chapter 1
December 10, 2012
Limits
namely, 0. To finish the proof we employ a series of algebraic manipulations followed by some
limit rules and the definition of e:
h
e −1
lim e h = lim (e h − 1 + 1) = lim
·h+1
← algebra
h
h→0
h→0
h→0
eh − 1
← limit rules
( lim h) + ( lim 1)
= lim
h
h→0
h→0
h→0
= (1)(0) + (1) = 1.
← Theorem 1.26
One technical point: In the second line of the above calculation we applied the product rule
for limits, which is only valid when the limits involved are known to exist. Therefore this proof
hinges on knowing that lim
h→0
eh −1
exists, which is a nontrivial fact that we will not prove here.
h
(b) To show that ln x is continuous on its domain, we must show that for all c ∈ (0, ∞) we have
lim ln x = ln c. We will use what we just proved about the continuity of e x . By the composition
x→c
rule for limits and the fact that e x is continuous everywhere, for c > 0 we have
lim ln x
lim e ln x = e x→c
x→c
.
Since ln x is the inverse of e x , we know that e ln x = x. Therefore for c > 0 we also have
lim e ln x = lim x = c = e ln c .
x→c
x→c
x
Since e is a one-to-one-function, putting these two calculations together we see that lim ln x
x→c
must be equal to ln c. (Note that once again, we are making an important assumption here,
that lim ln x is equal to some real number. If that number does not exist then we cannot apply
x→c
the composition rule for limits here.)
Using similar techniques, we can prove that trigonometric and inverse trigonometric functions are continuous on their domains. This again allows us to calculate limits of
trigonometric and inverse trigonometric functions at domain points by simple evaluation.
For example, cos x is defined at x = π , and this theorem says that the limit lim cos x is
x→π
equal to the value cos(π ) = −1. The theorem does not tell us how to calculate lim sec x,
x→π
however, because π is not in the domain of sec x.
THEOREM 1.28
Continuity of Trigonometric Functions
All trigonometric and inverse trigonometric functions are continuous on their domains.
Proof. We will prove Theorem 1.28 for the functions (a) sin x and (b) sin−1 x. Proofs for other basic
transcendental functions will be covered in Exercises 94–100.
(a) To show that sin x is continuous on its domain, we must show that lim sin x = sin c for all
x→c
c ∈ R. Following the same technique as for e x , we change variables with h = x − c and relate
to some limits that we already know. Recall that in the discussion after the Squeeze Theorem
we showed that lim sin x = 0 and lim cos x = 1. We thus have
x→0
x→0
lim sin x = lim sin(c + h)
x→c
← change variables
h→0
= lim (sin c cos h + sin h cos c)
h→0
= sin c lim cos h + cos c lim sin h
← sum identity for sine
= (sin c)(1) + (cos c)(0) = sin c.
← known limits
h→0
h→0
← sum and constant multiple rules
CONFIRMING PAGES
11:47
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
129
Limit Rules and Calculating Basic Limits
(b) Finally, to show that sin−1 xis continuous
on its domain [−1, 1] we use the fact that it is the
inverse of sin x restricted to −
π π
, with the same method we used for ln x. Before we do
,
2 2
this, a technical point: this proof will show that if lim sin−1 is equal to some real number, then
x→c
that real number must be sin−1 c. We will assume that lim sin−1 exists, a fact that is necessary
x→c
for the application of limit rules. By the composition rule for limits and the fact that sin x is
continuous everywhere, for c ∈ [−1, 1] we have
lim sin(sin−1 x) = sin( lim sin−1 x).
x→c
x→c
On the other hand, by properties of inverses, for c ∈ [−1, 1] we also have
lim sin(sin−1 x) = lim x = c = sin(sin−1 c).
x→c
x→c
π π
, we can put the preceding two calculations
Because sin x is a one-to-one function on − ,
2
together to conclude that lim sin−1 x = sin−1 c.
2
x→c
Delta-Epsilon Proofs of the Limit Rules
The limit rules seem almost obvious; for example, if f (x) approaches L and g(x) approaches
M as x → c, it is reasonable to expect that f (x) + g(x) approaches L + M as x → c. To prove
the limit rules in Theorem 1.20, however, we must appeal to the delta–epsilon definition
of limit.
Proof. (This proof requires material covered in optional Section 1.3.)
We will prove the (a) sum, (b) product, and (c) composition rules and leave the proofs of the
remaining rules to Exercises 91, 92, and 93.
(a) To prove the sum rule for limits, we must show that we can get f (x) + g(x) as close as we like
to L + M by choosing δ so that f (x) and g(x) are each half of that distance from L and M, as
illustrated in the graph that follows at the left. Given > 0, choose δ 1 to get f (x) within of L
2
and choose δ 2 to get g(x) within of M. Then for δ = min(δ 1 , δ 2 ) and x ∈ (c − δ, c) ∪ (c, c + δ),
2
we have
L − < f (x) < L +
and M − < g(x) < M + .
2
2
2
2
Adding these two double inequalities together, we get our desired conclusion:
(L + M) − < f (x) + g(x) < (L + M) + .
Given for f +g,
choose δ to get
2
Given for fg,
choose δ to get for f and g
for f and g
y
y
L M⑀
LM L M⑀
⑀
M2
M
L
L
LM f g
⑀
2
⑀
2
⑀
2
fg
M M L L g
f
c␦
c
c␦
x
g
f
cδ c cδ
CONFIRMING PAGES
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
130
Chapter 1
November 21, 2012
Limits
(b) The proof for the product rule for limits is similar: We will get f (x)g(x) as close as we like to
LM by choosing δ so that f (x) and g(x) are sufficiently close enough to L and M, respectively,
as illustrated in the preceding graph at the right. We will assume here that both L and M are
positive; the other cases are similar. Given > 0, choose > 0 sufficiently small so that < L,
< M, and (L + M) + ( )2 < . The reason for this odd choice of will become clear in
a moment. Now choose δ 1 and δ 2 to get f (x) and g(x) within of L and M, respectively. Then
for δ = min(δ 1 , δ 2 ) and x ∈ (c − δ, c) ∪ (c, c + δ), we have
L − < f (x) < L + and M − < g(x) < M + .
Since L and M are assumed to be positive, we can assume that δ has been chosen small enough
so that for x ∈ (c − δ, c) ∪ (c, c + δ) the values of f (x) and g(x) are also positive. Therefore,
multiplying our two double inequalities together gives
(L − )(M − ) < f (x)g(x) < (L + )(M + )
By our choice of it is easy to show that (L + )(M + ) < LM + and that LM − <
(L − )(M − ). Putting these two inequalities together with our double inequality for f (x)g(x),
we can say that for x ∈ (c − δ, c) ∪ (c, c + δ) we have
LM − < f (x)g(x) < LM + .
(c) To prove the composition rule for limits, let lim g(x) = L. We will show that we can get f (g(x)) as
x→c
close as we like to f (L) by choosing δ so that g(x) is sufficiently close to L. Since f is continuous
at L, we know that lim f (u) = L. Thus given > 0, we can choose δ > 0 so that whenever
u→L
u ∈ (L − δ , L) ∪ (L, L + δ ), we also have f (u) ∈ ( f (L) − , f (L) + ). In fact, if u = L, then
f (u) = f (L), so we can say a little bit more:
if u ∈ (L − δ , L + δ ), then f (u) ∈ ( f (L) − , f (L) + ).
Now, lim g(x) = L allows us to choose δ > 0 so that
x→c
if x ∈ (c − δ, c) ∪ (c, c + δ), then g(x) ∈ (L − δ , L + δ ).
Given for f (u), choose δ for u
Then given δ for u, choose δ for x
y
u
g
f
L δ
f(L) L
f(L)
f(L) L δ
L δ L L δ
u
cδ c
cδ
x
The two figures above illustrate our choices for δ and δ. Now given our choice of δ we can let
g(x) = u and string together the last two displayed implications above to conclude that
if x ∈ (c − δ, c) ∪ (c, c + δ), then f ( g(x)) ∈ ( f (L) − , f (L) + ).
Examples and Explorations
EXAMPLE 1
Calculating limits by using continuity and limit rules
Calculate the limits that follow, using only the continuity of linear and power functions
and the limit rules in Theorem 1.20. Cite each limit rule that you apply.
(a) lim (3x 2 − 2x + 1)
x→3
(b) lim (3x − 1)12
x→1
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
Limit Rules and Calculating Basic Limits
131
SOLUTION
(a) We apply the rules for limits of combinations of functions until we have reduced the
problem to limits of functions that we know to be continuous:
lim (3x 2 − 2x + 1) = lim 3x 2 − lim 2x + lim 1
x→3
x→3
x→3
x→3
← sum and difference rules
= 3 lim x 2 − 2 lim x + lim 1
← constant multiple rule
= 3(3 ) − 2(3) + 1 = 22.
← limits of continuous functions
x→3
2
x→3
x→3
(b) We could find this limit by multiplying out (3x − 1)12 and then applying the sum and
constant multiple rules as we did in part (a), but it is much faster to instead apply
the composition rule for limits. We know that the power function x12 is continuous
everywhere and, in particular, at lim (3x − 1) = 2. Therefore we have
x→1
lim (3x − 1)12 = (lim (3x − 1))12
x→1
← composition rule for limits
x→1
= ( 3(1) − 1 )12 = 212 = 4096.
EXAMPLE 2
← linear functions are continuous
Limits and continuity of piecewise-defined functions
Describe the continuity or discontinuity of each of the following piecewise-defined
functions at x = 1 by algebraically calculating the left, right, and two-sided limits
at x = 1:
x + 1, if x < 1
4 − x 2 , if x ≤ 1
(a) f (x) =
(b)
g(x)
=
2
3 − x , if x ≥ 1
x − 1, if x > 1
SOLUTION
(a) These two functions are the same ones we investigated graphically in Example 1 of
Section 1.4. Here, for each function we must compute the limit as x → 1 and compare
it with the value at x = 1. Since the piecewise-defined function f has a break point
at x = 1 we must compute its left and right limits separately. Because the component
functions x + 1 and 3 − x 2 are continuous, we can calculate the left and right limits
by evaluation:
lim f (x) = lim− (x + 1) = 1 + 1 = 2 and
x→1−
x→1
lim f (x) = lim+ (3 − x 2 ) = 3 − (1)2 = 2.
x→1+
x→1
Since the left and right limits both exist and are equal to 2, we have lim f (x) = 2, which
is equal to the value f (1) = 2. Therefore, f is continuous at x = 2.
x→1
(b) Once again we must calculate the left and right limits of g(x) separately. By continuity
of the component functions we have
lim g(x) = lim− (4 − x 2 ) = 4 − (1)2 = 3
x→1−
x→1
and
lim g(x) = lim+ (x − 1) = 1 − 1 = 0.
x→1+
x→1
Since the left and right limits both exist but are not equal as x → 1, the function g(x) has
a jump discontinuity at x = 1. Because lim− g(x) = 3 = g(1) but lim+ g(x) = 0 = g(1),
x→1
x→1
we can also say that g(x) is left continuous, but not right continuous, at x = 1.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
132
Chapter 1
EXAMPLE 3
December 10, 2012
Limits
Calculating limits
Calculate each of the following limits:
(a) lim
x→0
sin x +
π
2
2−x
x→2 4 − x 2
e 2x + e x − 2
ex − 1
x→0
(b) lim
SOLUTION
(a) f (x) = sin x +
π
2
(c) lim
is a combination of functions that are continuous on their do-
mains, and thus is continuous on its domain. Since x = 0 is in the domain of f we can
solve this limit by simple evaluation:
√
π
π
lim sin x +
= sin 0 +
= 1 = 1.
2
x→0
(b) The function f (x) =
2−x
4 − x2
2
is algebraic and thus continuous on its domain, but unfor-
tunately x = 2 is not in that domain. As x → 2, both the numerator and the denom0
inator approach 0, and therefore this limit is of the form , which is indeterminate.
0
This means that we don’t know at this point whether or not the limit exists or not, or
if it does, what it might be equal to. We can solve this indeterminacy by doing some
preliminary algebra; after cancellation we get a limit that is no longer indeterminate
and, in fact, that we can find by evaluation at x = 2:
2−x
lim
x→2 4 − x 2
= lim
2−x
x→2 (2 − x)(2 + x)
= lim
1
x→2 2 + x
1
4
= .
(c) As x → 0 both the numerator and the denominator approach zero, so this limit is of
0
the form , which is indeterminate. After factoring and cancelling we can resolve this
0
problem:
e 2x + e x − 2
(e x − 1)(e x + 2)
= lim
= lim (e x + 2) = e 0 + 2 = 3.
x
e −1
ex − 1
x→0
x→0
x→0
lim
EXAMPLE 4
Finding a Limit with the Squeeze Theorem
1
x
Use the Squeeze Theorem to find lim x 2 sin .
x→0
SOLUTION
The graph of f (x) = x 2 sin
1
x
f (x) = x 2 sin
1
x
follows at the left.
Squeezed between y = x 2 and y = −x 2
y
⫺0.4
y
0.10
0.10
0.05
0.05
⫺0.2
0.2
0.4
x
⫺0.4
⫺0.2
0.2
⫺0.05
⫺0.05
⫺0.10
⫺0.10
CONFIRMING PAGES
0.4
x
11:51
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
133
Limit Rules and Calculating Basic Limits
From this graph it seems reasonable that lim x 2 sin
x→0
1
x
= 0. We can use the Squeeze Theorem
to find this limit algebraically. The sine function has outputs that are always between −1
1
1
and 1. Therefore −1 < sin < 1, which means that −x 2 < x 2 sin < x 2 , as shown in the
x
x
graph at the right. We know that lim −x 2 = 0 and lim x 2 = 0. Therefore by the Squeeze
x→0
Theorem we must also have lim x 2 sin
x→0
EXAMPLE 5
1
x
x→0
= 0.
Using tables of values to approximate limits related to e
Use tables of values to approximate each of the limits that follow. In the second limit you
will have to use a calculator approximation of e to perform the calculations.
3h − 1
h
h→0
eh − 1
h
h→0
(a) lim
(c) lim (1 + h)1/ h
(b) lim
h→0
SOLUTION
(a) This limit is similar to the one in Theorem 1.26, but with the base e replaced by 3.
Limits of the form lim
h→0
bh − 1
h
converge to different numbers, depending on the base b,
but such a limit converges to 1 only for the number b = e. Thus we would expect that
3h − 1
should converge to some number other than 1. The following table of values
h→0 h
h
3 −1
suggests that this is indeed the case:
of
h
lim
h
−0.1
−0.01
−0.001
0
0.001
0.01
0.1
3 −1
h
1.040415
1.092600
1.098009
?
1.099216
1.104669
1.161232
h
If the pattern in the table continues, then we would expect this limit to converge to
some number between 1.098009 and 1.099216. The numbers in that range are close
to 1, but none of them are equal to 1. This is not unexpected, because the base in our
limit, which is 3, is close to the base e ≈ 2.71828 that would cause the limit we seek to
approach exactly 1.
(b) From Theorem 1.26 we know that lim
h→0
eh − 1
h
should approach 1. Let’s see if that is the
case. For small values of h approaching 0 we have
h
−0.1
−0.01
−0.001
0
0.001
0.01
0.1
e −1
h
0.951626
0.995017
0.999500
?
1.000500
1.005017
1.051709
h
As expected, this limit does seem to be approaching 1 as h approaches 0.
(c) We can approximate the limit in Definition 1.25 (and thus the value of e) by using a
table of approximate values of (1 + h)1/ h for small values of h:
h
(1 + h)
1/ h
−0.1
−0.01
−0.001
0
0.001
0.01
0.1
2.867972
2.731999
2.719642
?
2.716924
2.704814
2.593742
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
134
Chapter 1
November 21, 2012
Limits
Although (1 + h)1/ h does not seem to approach a very nice number as h → 0, it
appears that the limit does exist and is somewhere between 2.719642 and 2.716923. If
we evaluate (1 + h)1/ h at an extremely small value of h, say, h = 0.000001, then we can
get a relatively accurate approximation for the limit in Definition 1.25 and thus for the
number e:
e ≈ (1 + 0.000001)1/0.000001 ≈ 2.718280.
TEST YOUR
? UNDERSTANDING
The sum rule for limits says that the limit of a sum is the sum of the limits. In English,
what do the other limit rules say?
In the proof of the sum rule for limits, in order to get f (x) + g(x) within of L + M, how
close do we have to get f (x) and g(x) to L and M, respectively?
What is the rule for the limit of a constant? What is the rule for the limit of a constant
multiple of a function? How are these two rules different?
Why does it make sense that cancellation would be a valid operation when dealing with
a limit as x → c, even if what is being cancelled approaches zero as x → c?
In the Squeeze Theorem for limits, why do we require that the upper and lower func-
tions u(x) and l(x) have the same limit as x → c?
EXERCISES 1.5
Thinking Back
Values of transcendental functions: Without a calculator, find
each of the function values that follow. For some values the
answer may be undefined.
π .
If f (x) = csc x, find f (π ) and f
2
π .
If f (x) = tan2 x, find f (π ) and f
2
1
.
If f (x) = sin−1 x, find f (−1) and f
The δ– definition of limit: Write each limit statement that follows in terms of the δ– definition of limit. Then approximate
the largest value of δ corresponding to = 0.5, and illustrate
this choice of δ on a graph of f .
lim(3x − 2) = 4
lim(x 3 − 1) = 0
x→2
lim
x→1
x→1
x −1
=0
x+3
2
lim
x→0+
√
x+4 =2
2
If f (x) = tan−1
√
x, find f (1) and f (3).
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The limit of a difference of functions as
x → c is equal to the difference of the limits of those
functions as x → c, provided that all limits involved
exist.
(b) True or False: If f (x) is within 0.25 unit of 7 and g(x)
is within 0.25 unit of 2, then f (x) + g(x) is within
0.5 unit of 9.
(c) True or False: If f (x) is within 0.25 unit of 7 and g(x) is
within 0.25 unit of 2, then f (x)g(x) is within 0.5 unit
of 9.
(d) True or False: Every algebraic function f is continuous
at every real number x = c.
(e) True or False: Every power function f (x) = Ax k is continuous at the point x = 2.
(f) True or False: The function f (x) = sec x is continuous
π
at x = .
2
(g) True or False: The value of
to the limit of
(x − c)f (x)
at x = c is equal
(x − c)g(x)
f (x)
at x = c.
g(x)
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
(h) True or False: The limit of
(x − c)f (x)
as x → c is equal
(x − c)g(x)
f (x)
to the limit of
as x → c.
g(x)
Suppose f and g are functions such that lim f (x) = 5,
x→3
lim f (x) = 2, and lim g(x) = 4. Given this information, calcux→3
x→4
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Two limits that are initially in an indeterminate form
but can be solved with the Cancellation Theorem.
(b) Two limits that can be solved with the Squeeze
Theorem.
(c) Two limits that we do not yet know how to calculate.
3. State the constant multiple rule, sum rule, product rule,
quotient rule, and composition rule for limits.
4. Explain in your own words the types of functions whose
limits we can calculate with the limit rules in this section.
5. Explain why we can’t calculate every limit lim f (x) just by
135
Limit Rules and Calculating Basic Limits
late the limits that follow, if possible. If it is not possible with
the given information, explain why.
13. lim(2f (x) − 3g(x))
14. lim −2f (x)
15. lim f (x)
16. lim f (x)g(x)
x→3
x→7
17. lim
x→3
x→3
x→4
f (x) − 3
g(x)
18. lim f (g(x))
x→3
19. Graph the functions f (x) = x + 1 and g(x) =
x2 − 1
,
x−1
and show that they are equal everywhere except at one
point. Then show that f (x) and g(x) have different values,
but the same limit, at this point.
20. Graph the functions f (x) = 2 − x and g(x) =
x→c
4 − x2
,
x+2
evaluating f (x) at x = c. Support your argument with the
graph of a function f for which lim f (x) = f (c).
and show that they are equal everywhere except at one
point. Then show that f (x) and g(x) have different values,
but the same limit, at this point.
6. Find functions f and g and a real number c such that
lim f (x) + lim g(x) = lim( f (x) + g(x)). Does this example
21. In the Squeeze Theorem for limits, we require that l(x) ≤
f (x) ≤ u(x) for all x sufficiently close to c, but we do
not require this inequality to hold at the point x = c.
Why not?
22. Use a geometric argument and the Squeeze Theorem for
limits to argue that
x→c
x→c
x→c
x→c
contradict the sum rule for limits? Why or why not?
7. Find functions f and g and a real number c such that
(lim f (x))(lim g(x)) = lim( f (x)g(x)). Does this example
x→c
x→c
x→c
contradict the product rule for limits? Why or why not?
8. Write the constant multiple rule for limits in terms of
delta–epsilon statements.
9. Write the difference rule for limits in terms of delta–
epsilon statements.
10. Write the product rule for limits in terms of delta–epsilon
statements.
11. Explain how the algebraic function
√
f (x) = ( x + 1)3
is a combination of identity, constant, and power functions. Why does this mean that we can calculate limits of
this function at domain points by evaluation?
12. Explain how the algebraic function
f (x) =
lim sin θ = 0
θ →0−
for sufficiently small negative angles θ .
23. Use a geometric argument and the Squeeze Theorem for
limits to argue that
lim cos θ = 1
θ →0−
for sufficiently small negative angles θ.
24. In this exercise you will use a calculator to investigate the
number e.
(a) Make a table of values that describes the behavior of
the quantity (1 + h)1/ h as h → 0.
(b) Make a table of values that describes the behavior of
(x 2 + 1)(4 − 3x)
3x 2
the quantity
eh − 1
as h → 0.
h
(c) What do your tables of values have to do with Definition 1.25 and Theorem 1.26?
is a combination of identity, constant, and power functions. Why does this mean that we can calculate limits of
this function at domain points by evaluation?
Skills
Calculate the limits in Exercises 25–28, using only the continuity of linear and power functions and the limit rules. Cite
each limit rule that you apply.
x−1
25. lim 15(3 − 2x)
26. lim
x→1
x→−1 (x + 4)(x + 2)
3
28. lim
27. lim(3x + x 2 (2x + 1))
x→0 2x 2 − 4x + 1
x→3
Calculate each of the limits in Exercises 29–70.
29. lim(x 2 − 1)
x→0
30. lim(x − 1)(x + 1)(x + 5)
x→2
31.
3x
lim x 2 −
x→−1
x+2
4 + 2x
x+2
x2 − 1
35. lim
x→1 x − 1
x+3
37. lim
x→−3 3x 2 + 8x − 3
4x − 2
39. lim
x→1/2 6x 2 + x − 2
33. lim
x→2
32.
lim (3.1x 2 − 4x + 0.8)
x→1.7
4 + 2x
x+2
x2 − 1
36. lim
x→0 x − 1
3x 2 + 8x − 3
38. lim
x→−3
x+3
6x 2 + x − 2
40. lim
x→−2/3
3x + 2
34.
lim
x→−2
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
136
41.
Chapter 1
x−1
lim √
x→1+
x−1
x + x 2 − 2x 3
43. lim
x→1
x − x2
45. lim
(1 + h)2 − 1
h
47. lim
2 −3
4x
h→0
x
x→0
x
49. lim(3e 1.7x + 1)
x→4
51.
53.
55.
57.
59.
61.
63.
ex − 1
lim 2x
x→0 e
+ 2e x − 3
2x
lim
x→0 e x − 1
1
lim
x→π csc(x − π )
sin x
lim
x→0 tan x
x
lim x sin−1
x→1
2
1
lim
x→1− sin−1 x
(3 + h)2 − 32
lim
h→0
h
65. lim
h→0
67. lim
h→0
69. lim
h→0
(1 + h)3 − 13
h
1
−1
1+h
h
4
−1
(2 + h)2
h
November 21, 2012
Limits
42.
lim+
x→1
√
1− x
1−x
x + x 2 − 2x 3
44. lim
x→0
x − x2
(−1 + h)2 − 1
h→0
h
√
x
48. lim+ 2 − 4
71. f (x) =
72. f (x) =
46. lim
x→2
50.
52.
54.
56.
58.
60.
62.
64.
lim ln(1 +
x→0+
√
x)
ex − 1
lim 2x
x→1 e
+ 2e x − 3
x
e −1
lim
x→0
x
1 − cos(x − 1)
lim
x→1
x
cot x
lim
x→π/2 cos x
√
x
lim
√
x→3 tan−1
x
1
lim
x→0 sec−1 x
lim
h→0
66. lim
h→0
68. lim
h→0
70. lim
h→0
(2 + h)2 − 22
h
(−1 + h)3 − (−1)3
h
1
1
−
2+h
2
h
4
−4
(1 + h)2
h
Describe the intervals on which each function f in Exercises 71–78 is continuous. At each point where f fails to be
continuous, use limits to determine the type of discontinuity
and any left- or right-continuity.
73. f (x) =
74.
75.
76.
77.
78.
x 2 + 1, if x ≤ 0
1 − x, if x > 0
3x + 2, if x < −1
5 + 4x 3 , if x ≥ −1
x 2 − 3x − 1, if x = −2
3, if x = −2
⎧
2
⎪
⎨ 3x , if x < 2
4+x
f (x) =
⎪
⎩ x 2 − 2, if x ≥ 2
⎧ 2
x −1
⎪
⎪
⎨ x − 1 , if x < 1
f (x) =
0, if x = 1
⎪
⎪
⎩
3x − 1, if x > 1
⎧
⎨ x + 1, if x < 3
f (x) =
⎩ 2 2, if x = 3
x − 9, if x > 3
sin x, if x < π
f (x) =
cos x, if x ≥ π
⎧ x
⎨ 2 − 1, if x ≤ 0
f (x) = 4 x − 1
⎩
, if x > 0
2x − 1
Use the Squeeze Theorem to find each of the limits in Exercises 79–86. Explain exactly how the Squeeze Theorem applies
in each case.
1
1
80. lim x sin 2
79. lim x sin
x→0
x→0
x
x
1
1
82. lim sin x sin
81. lim(e x − 1) sin
x→0
x→0
x
x
1
1
2
84. lim(x − 1) cos
83. lim(x − 1) cos
x→1
x→1
x−1
x−1
1
1
86. lim x 2 tan−1
85. lim x tan−1
x→0
x→0
x
x
Applications
In Exercise 88 in Section 0.1, you constructed a piecewisedefined function from the 2000 Federal Tax Rate Schedule
that you will use in the next two problems. Specifically, you
found that a person who makes m dollars a year will pay T(m)
dollars in tax, given by the function
0.15m, if 0 ≤ m ≤ 26,250
3,937 + 0.28(m − 26,250), if 26,250 < m ≤ 63,550
14,381 + 0.31(m − 63,550), if 63,550 < m ≤ 132,600
35,787 + 0.36(m − 132,600), if 132,600 < m ≤ 288,350
91,857 + 0.396(m − 288,350), if m > 288,350.
87. Suppose you make $63,550 a year and pay taxes according
to the given formula.
(a) Calculate the value of T(63,550) and the limit of T(m)
as m approaches 63,550 from the left and from the
right.
(b) Use part (a) to argue that the function T(m) is
continuous at m = 63,550. What does this mean in
real-world terms?
88. Suppose you make $288,350 a year and pay taxes according to the given formula.
(a) Calculate the value of T(288,350) and the limit of
T(m) as m approaches 288,350 from the left and from
the right.
(b) Use part (a) to argue that the function T(m) is continuous at m = 288,350. What does this mean in realworld terms?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.5
Limit Rules and Calculating Basic Limits
137
Proofs
90. Use limit rules and the continuity of polynomial functions
to prove that every rational function is continuous on its
domain.
95. Use algebra, limit rules, and the continuity of e x to prove
that every exponential function of the form f (x) = Ab x is
continuous everywhere.
96. Use algebra, limit rules, and the continuity of ln x on
(0, ∞) to prove that every logarithmic function of the
form f (x) = logb x is continuous on (0, ∞).
91. Prove the constant multiple rule for
lim f (x) = L and k ∈ R, then lim kf (x) = kL.
97. In the reading we used the Squeeze Theorem to prove
that lim sin h = 0 and lim cos h = 1. Use these facts, the
89. Use limit rules and the continuity of power functions
to prove that every polynomial function is continuous
everywhere.
x→c
limits:
If
h→0
x→c
92. Prove the difference rule for limits by applying the sum
and constant multiple rules for limits.
93. Suppose that we know the reciprocal rule for limits: If
lim g(x) = M exists and is nonzero, then lim
x→c
x→c
1
1
= .
g(x)
M
This limit rule is tedious to prove and we do not include
it here. Use the reciprocal rule and the product rule for
limits to prove the quotient rule for limits.
94. Use algebra, limit rules, and the continuity of e x to prove
that every exponential function of the form f (x) = Ae kx is
continuous everywhere.
h→0
sum identity for cosine, and limit rules to prove that
f (x) = cos x is continuous everywhere.
98. Use the quotient rule for limits and the continuity of sin x
and cos x to prove that f (x) = tan x is continuous on its
domain.
99. Use the quotient rule for limits and the continuity of cos x
to prove that f (x) = sec x is continuous on its domain.
100. Use the composition rule for limits and the fact that tan x
is continuous on its domain to prove that tan−1 x is continuous everywhere.
Thinking Forward
Limits for Derivatives: In Chapter 2 we will define the
derivative of a function f at a point x = c to be the slope of
the line that points in the direction of the graph of f at x = c.
Algebraically, the derivative of f at c is given by the following
limit:
lim
h→0
f (c + h) − f (c)
.
h
0
Such limits are always of the indeterminate form , so we
0
must do algebra before we can resolve the limit.
Calculate the derivative of f (x) = x 2 at c = 0.
Calculate the derivative of f (x) = x 2 at c = 2.
Calculate the derivative of f (x) = x 2 at c = 4.
Sketch a graph of f (x) = x 2 , and sketch the lines that
point in the direction of the curve at (0, f (0)), (2, f (2)),
and (4, f (4)). Relate the slopes of these lines to the answers to the last three exercises.
Use the definition √
of the derivative to calculate the
derivative of f (x) = x at c = 4. At some point you will
need to multiply
by the
√
√ numerator and denominator
conjugate of 4 + h − 2, which is 4 + h + 2.
Use the definition of the derivative to calculate the
derivative of f (x) = x−1/2 at c = 4. As in the previous
calculation, you will need to multiply numerator and
denominator by a conjugate at some point.
Calculate the derivative of f (x) = e x at c = 0. At some
point you should need the characterization of e given
in Theorem 1.26.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
138
1.6
Chapter 1
November 21, 2012
Limits
INFINITE LIMITS AND INDETERMINATE FORMS
Calculating limits at infinity and infinite limits
Recognizing non-indeterminate forms and dealing with indeterminate forms
Special trigonometric limits
Infinite Limits
The utility of continuity is that it enables us to calculate limits by evaluation. However, we
1
can’t solve limits by evaluation if continuity fails. For example, consider the limit lim
.
The function
x→1 x − 1
1
x−1
is neither defined nor continuous at x = 1, so we cannot find its limit by
1
at x = 1, we would not get a real number, because the
evaluation. If we try to evaluate
x−1
denominator would be 0. In terms of limits, we say that
1
1
1
is of the form , and thus lim
is not a real number.
x
−
1
0
x
−
1
x→1
x→1
lim
Can we be more specific than just pointing out that limits of type
are not real numbers)? Consider the behavior of
1+
1
x−1
1
0
“do not exist” (i.e.,
as x approaches 1 from the right, as
shown in the table below. As x →
(see the first row of the table) the values of x − 1
approach 0+ (see the second row). The reciprocals of these values (see the third row) then
approach ∞:
x
1.1
1.01
1.001
1.0001
1.00001
→ 1+
x−1
0.1
0.01
0.001
0.0001
0.00001
→ 0+
1
x−1
10
100
1000
10,000
100,000
→ ∞
1
In symbols, this means that
→ ∞ as x → 1+ . In terms of limits we can express this
x−1
behavior from the right by saying that
lim
1
x→1+ x − 1
is of the form
1
1
, and thus lim+
= ∞.
0+
x→1 x − 1
We will prove a more general version of this statement in Theorem 1.29.
From the left, we have a similar situation: As x → 1− , we have x − 1 → 0− , so the
quantity
1
x−1
will remain negative as x → 1− , and as the magnitude of the denominator
x − 1 gets smaller and smaller, the magnitude of
1
x−1
gets larger and larger. In terms of
limits we can express this behavior from the left by saying that
lim
1
x→1− x − 1
is of the form
1
x→1 x − 1
We have now shown that lim
1
1
, and thus lim−
= −∞.
0−
x→1 x − 1
does not exist, and more specifically, that the limit is ∞
from the right and −∞ from the left.
If an expression approaches ∞ from both the right and the left, then we say that the
1
two-sided limit is ∞. For example, since
approaches ∞ from both the right and the
2
1
2
x→1 (x − 1)
left as x → 1, we would write lim
(x − 1)
is of the form
1
,
0+
1
2
x→1 (x − 1)
and thus lim
CONFIRMING PAGES
= ∞.
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
139
Infinite Limits and Indeterminate Forms
The theoretical basis for the discussion above is summarized in the following theorem:
THEOREM 1.29
Limits Whose Denominators Approach Zero from the Right or the Left
f (x)
x→c g(x)
is of the form
1
,
0+
then lim
f (x)
x→c g(x)
is of the form
1
,
0−
then lim
(a) If lim
(b) If lim
f (x)
x→c g(x)
= ∞.
f (x)
x→c g(x)
= −∞.
Theorem 1.29 also applies to one-sided limits and to limits as x → ∞ or as x → −∞. We
will prove only the case for limits from the right:
Proof. We will prove the case where lim+
x→c
f (x)
1
is of the form + . The other cases are similar; you
g(x)
0
will handle another in Exercise 87. Since f (x) → 1 as x → 0, it follows that for any 1 > 0, we can
find δ 1 > 0 such that
if c < x < c + δ 1 , then 1 − 1 < f (x) < 1 + 1 .
Similarly, since g(x) → 0+ as x → 0, it follows that for any 2 > 0, we can find δ 2 > 0 such that
if c < x < c + δ 2 , then 0 < g(x) < 0 + 2 .
f (x)
= ∞, take any M > 0. Choose δ to be the minimum of the δ 1 correg(x)
1
1
. With this choice of δ, we have
sponding to 1 = and the δ 2 corresponding to 2 =
2
2M
Now to prove that lim+
x→c
f (x)
1 − 1
if c < x < c + δ, then
>
=
g(x)
2
1
2 = M.
1
2M
Limits at Infinity
We have just seen that limits of the form
1
0+
and
1
0−
are always infinite. A sort of reverse
of this is also true, and is the subject of our next theorem: Limits of the form
1
∞
1
−∞
and
are always zero. This makes intuitive sense because the reciprocal of a number of large
magnitude is a number of small magnitude. Note that even though the expression “
1
”
∞
does not represent a real number, a limit of that form will be equal to a real number,
namely, 0.
THEOREM 1.30
Limits Whose Denominators Become Infinite Approach Zero
f (x)
x→∞ g(x)
is of the form
1
,
∞
f (x)
x→∞ g(x)
is of the form
1
,
−∞
(a) If lim
(b) If lim
f (x)
x→∞ g(x)
then lim
= 0.
f (x)
x→∞ g(x)
then lim
= 0.
1
1
→ 0; therefore lim
x+1
x→∞ x + 1
1
1
→ 0; therefore lim
x+1
x→−∞ x + 1
For example, as x → ∞ we have x + 1 → ∞ and thus
= 0.
Similarly, as x → −∞ we have x + 1 → −∞ and thus
= 0.
Theorem 1.30 also applies for limits as x → −∞ and as x → c, although we will not prove
that here.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
140
Chapter 1
November 21, 2012
Limits
Proof. We prove the first part and leave the second part to Exercise 88. Since f (x) → 1, it follows
that for any 1 > 0, we can find N1 > 0 such that
if x > N1 , then 1 − < f (x) < 1 + .
Similarly, since g(x) → ∞, it follows that for any M > 0, we can find N2 > 0 such that
if x > N2 , then g(x) > M.
f (x)
= 0, take any > 0. Choose N to be the maximum of the N1 correg(x)
2
sponding to 1 = 1 and the N2 corresponding to M = . With this choice of N, we have
Now to prove that lim
x→∞
if x > N, then 0 <
f (x)
1 + 1
2
<
= = .
2
g(x)
M
The next theorem lists the limits at infinity of some simple functions. You will prove a
selection of these limits in Exercises 90 and 91.
THEOREM 1.31
Limits of Some Basic Functions at Infinity
(a) If k > 0, then lim x k = ∞ and lim x−k = 0.
x→∞
x→∞
(b) If k > 0, then lim ekx = ∞ and lim e−kx = 0.
x→∞
x→∞
(c) lim ln x = ∞.
x→∞
(d) The functions sin x, cos x, tan x, sec x, csc x, and cot x all have periodic behavior
as x → ∞, and thus their limits as x → ∞ do not exist.
(e) lim tan−1 x =
x→∞
π
2
π
2
and lim tan−1 x = − .
x→−∞
Rather than memorizing these limits, it is better to remember the behavior on the right
side of the graphs of these basic functions, as in the following examples:
1
x
x2
y
y
x
ex
e−x
y
y
ln x
y
x
y
x
x
tan−1 x
sin x
y
x
x
?
x
We can say similar things about limits as x → −∞. In fact, in most cases limits as x → −∞
can be rewritten with algebra as limits as x → ∞. It is also helpful to remember the behavior on the left side of common graphs to determine limits as x → −∞. Although we
will not prove so in general here, the limit rules from Theorem 1.20 also apply when
x → ∞ and when x → −∞. You will prove this for the sum rule for limits in Exercise 89. For example, given that lim x−2 = 0 and that lim x−3 = 0, we can conclude that
lim (x−2 + x−3 ) = 0 + 0 = 0.
x→∞
x→∞
x→∞
As we will see in Example 4 and Exercise 85, polynomials behave like their leading
terms as x → ∞ and as x → −∞. For example, the function f (x) = 2x 3 − 5x − 1 will be
dominated by its leading term 2x 3 as x takes on larger and larger magnitudes. Therefore f
approaches ∞ as x → ∞, and approaches −∞ as x → −∞. Since rational functions are by
definition quotients of polynomial functions, we can use what we know about the global
behavior of polynomials to determine the global behavior of rational functions:
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
THEOREM 1.32
141
Infinite Limits and Indeterminate Forms
Horizontal Asymptote Theorem for Rational Functions
p(x)
If f (x) =
is a rational function in which the polynomials p(x) and q(x) have leading
q(x)
terms a n x n and b m x m , respectively, then
(a) if n < m, then the graph of y = f (x) has a horizontal asymptote at y = 0.
(b) if n = m, then the graph of y = f (x) has a horizontal asymptote at y =
an
.
bm
(c) if n > m, then the graph of y = f (x) does not have a horizontal asymptote.
Indeterminate and Non-Indeterminate Forms
If a limit is indeterminate, then we cannot initially say whether or not it exists—or if it
exists, what real number it is equal to. Many indeterminate limits can be resolved with
algebra such as cancellation or factoring. For example, the four limits that follow are all
0
initially of the indeterminate form . After some simple algebra, we see that three limits
0
exist and one limit becomes infinite. The three that exist are each equal to different real
numbers.
(x − 1)2
x−1
0
= lim
= =0
1
x→1 x − 1
x→1 1
lim
lim
x−1
x→1 x − 1
= lim 1 = 1
x→1
lim
x−1
x→1 (x − 1)3
lim
x−1
x→1 3(x − 1)
1
= lim
x→1 (x − 1)2
= lim
1
x→1 3
=
=∞
1
3
When a limit is indeterminate, it is essentially because there is a “fight” going on between
0
two parts of the limit. For example, limits of the form approach different things depend0
ing on whether the numerator or the denominator approaches 0 faster. If the numerator
does, then it “wins” the fight and the limit is equal to zero. If the denominator approaches
0 faster, then the limit will become infinite. If the numerator and denominator are balanced
appropriately, then they cancel each other out and the limit will approach a nonzero real
number. Over time you will develop an intuition for what types of expressions are likely to
win such fights, and the examples at the end of this section illustrate algebraic techniques
for resolving such indeterminacies.
The following theorem identifies seven common indeterminate forms for limits:
THEOREM 1.33
Indeterminate Forms for Limits
Each of the following is an indeterminate form, meaning that a limit in one of these
forms may or may not exist, depending on the situation:
0
0
∞
∞
0·∞
∞−∞
00
1∞
∞0
To prove this theorem, for each indeterminate form we need only exhibit an example
of a limit of that form that exists and an example of a limit of that form that does not,
0
as we did before for limits of the indeterminate form . Identifying such examples for
0
the remaining six forms is left to you in Exercises 15–21. Limits that have the first four
types of indeterminate forms listed in Theorem 1.33 can often be solved after simple factoring or cancelling. Limits of the remaining three types pose more of a challenge; see
Section 3.6.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
142
Chapter 1
CAUTION
November 21, 2012
Limits
0
Of course, the indeterminate expression “ ” is not a real number and we cannot actually
0
divide the number 0 by the number 0. Theorem 1.33 tells us that limits of this form are
indeterminate, and thus cannot be determined until we somehow rewrite or re-examine
the limit, perhaps by factoring, cancelling, or some other method. Note that this is very
different than saying that a limit “does not exist.”
We have already seen that a limit of the form
1
0+
1
∞
is equal to zero and a limit of the form
is infinite. The following limit forms also always either approach 0 or become infinite
and thus are not indeterminate:
THEOREM 1.34
Non-Indeterminate Forms for Limits
(a) A limit in any of these forms must be equal to 0:
1
∞
0
∞
0
1
0∞
01
(b) A limit in any of these forms must be ∞:
1
0+
∞
0+
∞
1
∞+∞
∞·∞
∞∞
∞1
We are not suggesting that you should memorize Theorem 1.34. Each of the limit forms
in the theorem can be easily determined by investigation. We will not give a formal
proof of this theorem, but rather, we can argue that in each case there is no “fight”; it
is clear what limits of each form must approach. For example, in a limit of the form
0
,
∞
the numerator approaches zero, making the quotient smaller and smaller; at the same
time the denominator grows without bound, which also makes the quotient smaller and
smaller. Thus for the form
0
,
∞
the behavior of both the numerator and the denominator
causes the limit to approach 0. You will see some of the other non-indeterminate forms in
Exercises 11–14.
Special Trigonometric Limits
Certain indeterminate limits can be reduced with algebra to two specific trigonometric limits. These limits expand the library of limits that we can compute and will be vital tools for
determining the derivatives of sine and cosine in the next chapter.
THEOREM 1.35
Two Useful Trigonometric Limits
(a) lim
θ→0
sin θ
=1
θ
(b) lim
θ→0
1 − cos θ
=0
θ
Notice that the limits in this theorem give us a way to determine a number of related limits
0
that are initially of the form . For example, we can use the first limit to show that
0
lim
x→0
sin 2x
sin 2x
= lim
(2) = 1(2) = 2,
x
x→0 2x
since as x → 0, we also have 2x → 0. As another example, since x − π → 0 when x → π ,
we have
lim
x→π
sin(x − π )
= 1.
x−π
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
1.6
143
Infinite Limits and Indeterminate Forms
Proof. We can intuitively see that the two special trigonometric limits make sense based on the
following figures of portions of the unit circle, with angles measured in radians:
1 − cos θ much smaller than θ
sin θ ≈ θ
y
y
θ
tan θ
sin θ
θ
θ0
θ
x
x
cos θ
1
1 cos θ
For small positive values of θ, the picture on the left suggests that sin θ ≈ θ and therefore that
sinθ
≈ 1. Similarly, for small positive θ, the picture on the right suggests that 1 − cos θ approaches
θ
1 − cosθ
0 much faster than θ and therefore that
≈ 0.
θ
More formally, from the earlier figure on the left, we can see that sin θ ≤ θ ≤ tan θ . Dividing
all expressions in this chain of inequalitites by sin θ and using the fact that tan θ =
1≤
sinθ
, we have
cosθ
1
θ
≤
.
sin θ
cos θ
Since all of the expressions in this chain of inequalities are positive, we can take reciprocals to
obtain:
sin θ
cos θ ≤
≤ 1.
θ
sinθ
= 1. The
θ
Now by applying the Squeeze Theorem to these two inequalities we have lim+
θ →0
argument for θ → 0− is similar, with a picture in the fourth quadrant.
The second trigonometric limit can be proved from the first by using the double-angle formula
cos 2θ = 1 − 2 sin2 θ ; see Exercise 94.
1 − cosx
sinx
The two figures that follow show the functions y =
and y =
. Notice that
x
x
neither function is defined at x = 0, but both approach a specific real-number value as
x → 0.
lim
x→0
sinx
=1
x
lim
x→0
1 − cosx
=0
x
y
y
1
1
3π
3π
π
π
3π
π
3π
π
x
1
Examples and Explorations
EXAMPLE 1
Calculating limits to determine horizontal asymptotes
Use limits to find any horizontal asymptotes of the following functions:
(a) f (x) =
x
x−1
(b) g(x) =
x2 − 1
x 3 − 4x
(c) h(x) =
sin x
x
CONFIRMING PAGES
x
11:51
TKmaster2010
WHF00153/FREE087-Taalman
144
Chapter 1
November 21, 2012
Limits
SOLUTION
(a) To find horizontal asymptotes we must examine limits as x → ±∞. As x → ∞, we
have x → ∞ and x − 1 → ∞, and therefore
lim
x
x→∞ x − 1
is of the indeterminate form
∞
.
∞
∞
Limits of the indeterminate form
can often be resolved by dividing the numerator
∞
and denominator by the highest power of x that appears, as follows:
x
x
1/2x
1
1
lim
= lim
=
= 1.
= lim
x→∞ x − 1
x→∞ x − 1
1/2x
In a similar fashion we can show that
x→∞ 1 − (1/x)
x
lim
x→−∞ x − 1
1−0
= 1. Therefore f (x) =
two-sided horizontal asymptote at y = 1.
x
x−1
has a
(b) As x → ∞ we have x 2 − 1 → ∞ in the numerator and x 3 → ∞ and 4x → ∞ in the
denominator. Since ∞ − ∞ is an indeterminate form, we cannot even be certain what
the denominator x 3 − 4x approaches at this point. To resolve this limit we will once
again divide the numerator and denominator by the highest power of x:
x2 − 1
x 2 − 1 1/x 3
1/x − 1/x 3
0−0
0
lim 3
= = 0.
= lim 3
=
=
lim
1/x 3
1−0
1
x→∞ x − 4x
x→∞ x − 4x
x→∞ 1 − (4/x 2 )
x2 − 1
x→−∞ x 3 − 4x
is equal to 0. Therefore g(x) =
Similarly, lim
asymptote at y = 0.
x2 − 1
x 3 − 4x
has a two-sided horizontal
sinx
as x → ∞, we see that the numerator sin x oscillates between
(c) Looking at h(x) =
x
−1 and 1 while the denominator x gets infinitely large. A bounded quantity divided
by a quantity that increases without bound must approach zero; in other words, as
x → ∞ we have
sin x
bounded
→
→ 0.
x
∞
The same is true as x → −∞, and thus lim
x→∞
Therefore the function h(x) =
CHECKING
THE ANSWER
sinx
x
sinx
x
and lim
x→−∞
sinx
x
are both equal to 0.
has a two-sided horizontal asymptote at y = 0.
We can use calculator graphs to verify the horizontal asymptotes that we just found. These
graphs also provide verification for the vertical asymptotes in the next example.
f has horizontal asymptote at y = 1
g has horizontal asymptote at y = 0
5
h has horizontal asymptote at y = 0
3
2
4
6
1.2
6
30
3
EXAMPLE 2
30
0.2
3
Calculating limits to determine vertical asymptotes
Use limits to describe any vertical asymptotes of the following functions:
(a) f (x) =
x
x−1
(b) g(x) =
x2 − 1
x 3 − 4x
(c) h(x) =
sin x
x
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
Infinite Limits and Indeterminate Forms
145
SOLUTION
(a) From the formula for f (x), we see that x = 1 is the only serious candidate for a vertical
asymptote. As x → 1 we have x − 1 → 0 and therefore
lim
x
x→1 x − 1
1
0
is of the form .
x
has a vertical asymptote at x = 1. If we want to describe
This tells us that f (x) =
x−1
the behavior of f near this asymptote more precisely, we can calculate the right and left
limits separately. As x → 1− we have x − 1 → 0− , and as x → 1+ we have x − 1 → 0+ ;
therefore
x
x
lim−
= −∞ and lim+
= ∞.
x→1
x−1
x→1
x−1
This means that the vertical asymptote at x = 1 is downward-pointing on the left and
upward-pointing on the right; see the leftmost graph from the previous “Checking the
Answer” figures.
(b) The function g(x) factors as
x2 − 1
(x − 1)(x + 1)
=
.
3
x − 4x
x(x − 2)(x + 2)
The values of x that cause the denominator of this quotient to approach zero are x = 0,
x = 2, and x = −2. None of these values cause the numerator to approach zero, so in
each case we will get a limit that is either ∞ or −∞ from the left and/or the right. In
any case, we know that g(x) has vertical asymptotes at x = 0, x = −2, and x = 2. If we
want to know the precise behavior of g(x) at one of these vertical asymptotes, we can
look from the left and the right. For example, as x → 2− we have x(x − 2)(x + 2) →
2(0− )(4) → 0− , and as x → 2+ we have x(x − 2)(x + 2) → 2(0+ )(4) → 0+ . Therefore
the left and right limits at x = 2 are
lim−
x2 − 1
(x − 1)(x + 1)
= lim−
= −∞
3
x − 4x x→2 x(x − 2)(x + 2)
lim+
x2 − 1
(x − 1)(x + 1)
= lim
= ∞.
x 3 − 4x x→2+ x(x − 2)(x + 2)
x→2
and
x→2
Notice that in these calculations we kept track only of the left/right “±” directions
when we encountered 0. This is because whether a multiplicative factor in the limit
is approaching, say, 1+ or 1− , will not affect the overall sign. On the other hand, the
difference between a factor of 0+ and 0− does affect the sign, which in turn will determine whether the limit approaches ∞ or −∞. Notice also that we used the factored
form of g(x) to determine the sign of infinity in each case, since it would have been
difficult to determine the sign of the denominator as x → 2+ and as x → 2− in the
unfactored expression for g(x).
sinx
(c) The denominator of h(x) =
is zero only when x = 0, so x = 0 is the only candidate
x
for a vertical asymptote. However, at x = 0 we have one of our special limits from
Theorem 1.35:
sin x
lim
= 1.
x→0
x
Therefore h(x) does not have a vertical asymptote at x = 0. Instead, the graph has a
hole, since lim h(x) = 1 exists but h(0) does not exist; see the third graph in the Checkx→0
ing the Answer discussion before this example.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
146
Chapter 1
EXAMPLE 3
December 10, 2012
Limits
Indeterminate forms
Determine whether each of the limits that follows is initially in indeterminate form or nonindetermine form. Then calculate each limit.
3x − 2x
3 2x
2/3
3/4
(a) lim
1
+
(b)
lim
(x
−
x
)
(c)
lim
x
x→∞ 1 + 4
x→∞
x→∞
x
SOLUTION
(a) By Theorem 1.31, as x → ∞ the expressions 3 x , 2 x , and 4 x all approach ∞. Therefore the limit in question is an indeterminate form. An analog of the method of
dividing by the highest power works in this case, but this time we divide numerator
and denominator by the exponential function with the largest base:
3x − 2x
3 x − 2 x 1/4 x
0−0
0
(3/4)x − (1/2)x
lim
=
lim
=
= = 0.
= lim
x
x
x
x→∞ 1 + 4
x→∞ 1 + 4
x→∞
1/4
0+1
1
(1/4)x + 1
(b) Since
2
3
3
4
and
are positive powers, as x → ∞ we have x 2/3 → ∞ and x 3/4 → ∞.
Therefore the limit in question is of the indeterminate form ∞ − ∞. Limits of this
form can often be resolved by factoring. We have
x 2/3 − x 3/4 = x 2/3 (1 − x 1/12 ),
and as x → ∞ we have x 2/3 → ∞ and (1 − x 1/12 ) → −∞. Therefore the limit in
question is now of the form (∞)(−∞) which means that
lim (x 2/3 − x 3/4 ) = lim x 2/3 (1 − x 1/12 ) = −∞.
x→∞
x→∞
(c) As x → ∞, the base 1 +
3
x
approaches 1 + 0 = 1 and the exponent 2x approaches
∞. Therefore this limit is of the form 1∞ , which is indeterminate. Fortunately, with a
substitution we can rewrite the limit in such a way that allows us to apply the definition
3
of e from Definition 1.25. Let h = . Then as x → ∞, we have h → 0+ . Using this
3
h
x
relationship and the fact that x = , we have
3 2x
= lim (1 + h)2(3/ h) = ( lim (1 + h)1/ h )6 = e 6 .
lim 1 +
x→∞
CHECKING
THE ANSWER
x
h→0+
h→0+
We can use intuition to verify that these answers seem reasonable. Remember that each
time a limit has an indeterminate form, two parts of the limit are fighting against each
other. In part (a) of the preceding example,
3x − 2x
1 + 4x
approaches an indeterminate form as
is the exponential function with the largest base in the expression, it is
x → ∞. Since
reasonable to expect that it will dominate the expression as x → ∞, dragging the whole
limit down to zero.
4x
In part (b) we saw that lim x 2/3 − x 3/4 was of the indeterminate form ∞ − ∞. Since
3
4
2
3
x→∞
> , x 3/4 should approach ∞ faster than x 2/3 does. Thus it makes sense to expect that
x 3/4 should win the battle and x 2/3 − x 3/4 should eventually approach −∞.
It is difficult to use intuition to verify the limit in part (c), except that we might expect in this case that as x → ∞ the 1 and the ∞ are balanced in the indeterminate form
1∞ , due to the fact that each involve a single power of x. This might lead us to suspect that
the answer is neither 1, nor ∞, but rather some number in between.
CONFIRMING PAGES
12:19
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
EXAMPLE 4
147
Infinite Limits and Indeterminate Forms
The global behavior of a polynomial is determined by its leading term
Use limits to show that the polynomial f (x) = x 4 − x 3 − 11x 2 + 9x + 18 behaves like its
leading term x 4 as x → ∞ and x → −∞. Then use graphs to compare the graph of f with
the graph of y = x 4 in different graphing windows.
SOLUTION
It is not immediately obvious how to calculate this limit, because, as x → ∞, the terms x 4
and 9x approach ∞ while the terms −x 3 and −11x 2 approach −∞. Therefore lim x 4 − x 3 −
x→∞
11x 2 + 9x + 18 is indeterminate.
However, with some simple algebra we can change this sum and difference of infinities
into a product that is easier to work with. Specifically, we can factor out the largest power
of x:
1
11
9
18
lim (x 4 − x 3 − 11x 2 + 9x + 18) = lim x 4 1 − − 2 + 3 + 4 .
x→∞
x
x→∞
x
x
x
Since as x → ∞ we have x 4 → ∞ and the remainder of the expression approaching
1 − 0 − 0 + 0 + 0 = 1, we can say that the limit is equal to ∞.
Similarly, the limit as x → −∞ is also ∞. Notice that the only term which ended up
being relevant in the limit calculation was the leading term. The figures that follow show
the function f (x) = x 4 − x 3 − 11x 2 + 9x + 18 in blue and its leading term y = x 4 in red,
in three different viewing windows. The more we enlarge the graphing window, the more
the graph of the function y = f (x) looks like the graph of y = x 4 .
f (x) and x 4 on [−3, 3]
f (x) and x 4 on [−6, 6]
y
f (x) and x 4 on [−8, 8]
y
y
500
20
3000
400
10
300
1
3 2 1
2
3
x
200
6 4 2
100
20
TEST YOUR
1000
100
10
? UNDERSTANDING
2000
2
4
6
x
8 6 4 2
2
4
6
8
x
In terms of large and small numbers, why does it make intuitive sense that limits of the
form
1
0+
must always equal ∞?
If a limit is of the form
right limits separately?
1
0
as x → c, why should we examine the corresponding left and
Why does it make sense that limits of the form
to zero?
1
∞
and of the form
Why does it make intuitive sense that limits of the form
1
−∞
are always equal
0
0
are indeterminate? What
“fight” is happening between the numerator and the denominator? What will happen
if the numerator “wins”? The denominator? If there is a tie?
Can a function have two different horizontal asymptotes? Can you think of a specific
example?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
148
Chapter 1
November 21, 2012
Limits
EXERCISES 1.6
Thinking Back
Behavior of transcendental functions: Determine whether each
function approaches 0, approaches a nonzero real number, or
becomes infinite as x approaches each indicated value.
f (x) = csc x, with x → 0 and x → π.
f (x) = tan2 x, with x → 0 and x → π .
−1
f (x) = sin
f (x) = tan−1
x, with x → 0 and x → 1.
√
x, with x → 0 and x → 3.
The definition of infinite limits and limits at infinity: Write each
limit statement that follows in terms of the formal definition
of limit. Then approximate the largest value of δ or N corresponding to = 0.5 or M = 100, as appropriate, and illustrate
this choice of δ or N on a graph of f .
lim
x→∞
lim
x→∞
2x
=2
x−1
1
=∞
x2 − 4
1
lim
=0
x→−∞ x
√
x−1=∞
lim
x→2+
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If f (x) → 0+ , then
1
→ ∞.
f (x)
(b) True or False: If f (x) → ∞+ , then
1
→ 0+ .
f (x)
(c) True or False: If a limit initially has an indeterminate
form, then it can never be solved.
(d) True or False: A limit “does not exist” if there is no real
number that it approaches.
(e) True or False: As limit forms, ∞2 → ∞.
(f) True or False: As limit forms, 2∞ → ∞.
(g) True or False: As limit forms, ∞ − ∞ → 0.
(h) True or False: The limit of a function f as x → c is
always equal to the value f (c), provided that f (c)
exists.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A limit of the form
and −∞ as x → c− .
1
that approaches ∞ as x → c+
0
(b) Two limits that can be solved with the special trigonometric limits from Theorem 1.35.
(c) Formulas for three functions that are discontinuous
at x = 3: one removable, one jump, and one infinite
discontinuity.
In Exercises 3–6, lim f (x) = L and lim g(x) = M for some
x→c
x→c
real numbers L and M. What, if anything, can you say about
lim
x→c
7. Determine which of the given forms are indeterminate.
For each form that is not indeterminate, describe the behavior of a limit of that form.
∞+∞
3. L = 0 and M = 0
4. L = 0 and M = 0
5. L = 0 and M = 0
6. L = 0 and M = 0
∞+1
0+∞
8. Determine which of the given forms are indeterminate.
For each form that is not indeterminate, describe the
behavior of a limit of that form.
∞·∞
0·∞
5·∞
5·0
0·0
9. Determine which of the given forms are indeterminate.
For each form that is not indeterminate, describe the
behavior of a limit of that form.
0
0
0
∞
∞
0
1
0
∞
1
0
1
∞
∞
1
∞
10. Determine which of the given forms are indeterminate.
For each form that is not indeterminate, describe the
behavior of a limit of that form.
01
00
0∞
1∞
∞1
∞0
∞∞
11. Describe in terms of large and small numbers why it
makes intuitive sense that limits of the form (a)
(b)
0
0
, and (c) must equal 0.
∞
1
1
,
∞
12. Describe in terms of large and small numbers why it
makes intuitive sense that limits of the form (a) 0∞ and
(b) 01 must equal 0.
13. Describe in terms of large and small numbers why it
makes intuitive sense that limits of the form (a)
(b)
f (x)
in each case?
g(x)
∞−∞
∞
∞
, and (c)
must be infinite.
0+
1
1
,
0+
14. Describe in terms of large and small numbers why it
makes intuitive sense that limits of the form (a) ∞ + ∞,
(b) ∞ · ∞, (c) ∞∞ , and (d) ∞1 must be infinite.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
To prove that the limit forms in Theorem 1.33 are indeterminate, we need only list explicit examples of limits that do and
do not exist for each form. Do so for each of the limit forms
from Exercises 15–21. For the last three forms you may want
to experiment with a graphing utility to find your examples.
15.
0
that approaches (a) 0, (b) 2, (c) ∞.
0
16.
∞
that approaches (a) 1, (b) 6, (c) ∞.
∞
149
Infinite Limits and Indeterminate Forms
21. ∞0 that approaches (a) 1, (b) 2, (c) ∞.
22. Find the equation of a rational function that could have
the graph shown. Take into account roots, holes, and
vertical and horizontal asymptotes when constructing
your function.
y
10
17. 0 · ∞ that approaches (a) 0 (b) 1, (c) ∞.
2
18. ∞ − ∞ that approaches (a) 0, (b) 5, (c) ∞.
2 1
1
x
2
19. 00 that approaches (a) 1, (b) 0, (c) ∞.
10
20. 1∞ that approaches (a) 1, (b) e, (c) ∞.
Skills
Find the roots, discontinuities, and horizontal and vertical
asymptotes of the functions in Exercises 23–34. Support your
answers by explicitly computing any relevant limits.
x 2 − 2x − 3
x−3
(x+1)(x−2)
25. f (x) =
(x−2)(x+2)
(x+1)(x−2)
27. f (x) =
(x−2)2 (x+2)
26. f (x) =
2
29. f (x) =
4 + e −2x
1
30. f (x) =
2 + 3x
2x − 4x
31. f (x) =
3x
4 x − 6(2 x ) + 5
32. f (x) =
1 − 2x
23. f (x) =
−1
33. f (x) = tan (3x) + 1
24. f (x) =
2x 2 − 1
− 2x + 1
x2
2
(x+1)(x−2)
(x−2)(x+2)
(x+1)(x−2)
28. f (x) =
(x−2)(x+2)2
1
34. f (x) =
tan−1 x
51.
53.
36.
lim 2x−4/3
38.
x→0
37.
39.
41.
x→∞
√
lim ( x − x)
x→∞
lim (−3x 5 + 4x + 11)
x→∞
40.
42.
x−3
− x−1
57.
lim
x→∞
4x − 3x
5x
3x − 5x
59. lim
x→−∞
4x
61.
63.
65.
lim
x→∞
2e 1.5x
3e 2x + e 1.5x
54.
56.
58.
60.
62.
lim ln(x 2 − 9)
64.
lim ( ln x − ln(2x + 1))
66.
x→3+
2
x→∞
lim (x−1/3 − x−1/2 )
x→∞
x−3
− x−1
lim+
x2
lim
x 7/2 − x 8/3
x2
lim
2 x − 4−x
3x
lim
4(3 x )
2 + 3x
x→0
x→∞
x→∞
x→∞
1 − 5e 2x
3e x + 4e 2x
1
lim+ ln
lim
x→∞
x
x→0
lim ( ln 3x − ln 2x)
x→∞
67. lim
1 − cos 2x
7x
68. lim
sin 3x
5x
lim −5x 3/5
69. lim
x
1 − cos x
70. lim
3 sin x + x
x
71. lim
sin2 3x
x3 − x
72. lim
sin(3x 2 )
x3 − x
74. lim
x 2 cot x
sin x
x→−∞
x→0
x→0
lim (x 4 − x 5 )
x→−∞
lim (5 − 2x + 3x 3 )
x→0
x→−∞
x+1
44. lim
x→2 (x − 2)2
x2 + 1
x→0 x(x − 1)
x
47. lim 2
x→0 x − x
x+4
x→−4 x 2 + 8x + 16
x−1
48. lim 2
x→1 x − 2x + 1
(3x + 1)2 (x − 1)
49. lim
x→∞
1 − x3
1 − 2x 2
50. lim
x→−∞ (3 − x)(3 + 4x)
46.
x2
52.
lim 2x−3/4
x→0
x 2 + 8x + 16
43. lim
x→−4 (x + 4)2 (x + 1)
45. lim
lim
x→∞
x 7/2 − x 8/3
55. lim+
x→0
x2
Calculate each limit in Exercises 35–80.
35. lim −4x−3
lim (x−1/3 − x−1/2 )
x→0+
lim
x→0
x→0
x→0
2
73.
lim+
x→0
x csc 3x
1 − cos 2x
x→0
sec x tan x
75. lim
x→0
x
76. lim 3x 2 cot2 x
77. lim(1 + x)2/x
78. lim(1 + 2x)3/x
x→0
79.
lim
x→∞
1+
1
x
x→0
x→0
3x
80.
lim
x→∞
CONFIRMING PAGES
1−
5
x
x
15:42
TKmaster2010
WHF00153/FREE087-Taalman
150
Chapter 1
December 10, 2012
Limits
Applications
81. In 1960, H. von Foerster suggested that the human
population could be measured by the function
P(t) =
179 × 109
.
(2027 − t)0.99
The time t is measured in years, where t = 1 corresponds
to the year 1 a.d., t = 1973 corresponds to the year 1973
a.d., and so on. (We saw this “doomsday model” for population in Problem 77 of Section 1.1, on page 89.) Use limit
techniques to calculate lim − P(t). What does this limit
t→2027
mean in real-world terms?
82. Suppose instead we consider the population model
Q(t) =
44 × 1010
,
1 + (2027 − t)4/3
with t measured in years as in the previous problem.
(a) Use limit techniques to calculate lim Q(t). What does
t→∞
this limit mean in real–world terms? What happens
in this model in the year 2027?
(b) Use calculator graphs to compare the population
models in this exercise with those in the previous
exercise. Describe the long–term population growth
scenarios that are suggested by these models.
83. Consider a mass hanging from the ceiling at the end of
a spring. If you pull down on the mass and let go, it will
oscillate up and down according to the equation
s(t) = A sin
k
t + B cos
m
k
t ,
m
where s(t) is the distance of the mass from its equilibrium
position, m is the mass of the bob on the end of the spring,
and k is a “spring coefficient” that measures how tight or
stiff the spring is. The constants A and B depend on initial
conditions—specifically, how far you pull down the mass
(s0 ) and the velocity at which you release the mass (v0 ).
This equation does not take into effect any friction due to
air resistance.
(a) Determine whether or not the limit of s(t) as t → ∞
exists. What does this say about the long-term
behavior of the mass on the end of the spring?
(b) Explain how this limit relates to the fact that the equation for s(t) does not take friction due to air resistance
into account.
(c) Suppose the bob at the end of the spring has a mass
of 2 grams and that the coefficient for the spring is
k = 9. Suppose also
√ that the spring is released in such
a way that A = 2 and B = 2. Use a graphing utility
to graph the function s(t) that describes the distance
of the mass from its equilibrium position. Use your
graph to support your answer to part (a).
84. In the previous exercise we gave an equation describing spring motion without air resistance. If we take into
account friction due to air resistance, the mass will oscillate up and down according to the equation
4km − f 2
t
s(t) = e (−f / 2m)t A sin
+ B cos
2m
s0
v0
,
where m, k, A, and B are the constants described in
Problem 83 and f is a positive “friction coefficient” that
measures the amount of friction due to air resistance.
(a) Find the limit of s(t) as t → ∞. What does this say
about the long-term behavior of the mass on the end
of the spring?
(b) Explain how this limit relates to the fact that the new
equation for s(t) does take friction due to air resistance
into account.
(c) Suppose the bob at the end of the spring has a mass
of 2 grams, the coefficient for the spring is k = 9, and
the friction coefficient is f = 6. Suppose also that the
spring is released in such a way that A = 4 and B = 2.
Use a graphing utility to graph the function s(t) that
describes the distance of the mass from its equilibrium position. Use your graph to support your answer
to part (a).
spring coefficient, k
s(t)
4km − f 2
t
2m
mass, m
CONFIRMING PAGES
12:19
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
1.6
Infinite Limits and Indeterminate Forms
151
Proofs
85. Use limits to prove that the limits of a polynomial f (x) =
a n x n + a n−1 x n−1 + a1 x + a0 are the same as the limits of
its leading term a n x n as x → ∞ and as x → −∞. (Hint:
Show that lim f (x) is equal to lim a n x n by factoring out
x→∞
x→∞
a n x n from f (x).)
86. Use limit techniques to prove that a rational function
p(x)
f (x) =
will have
q(x)
(a) a horizontal asymptote at y = 0, if the degree of p(x)
is less than the degree of q(x);
a
(b) a horizontal asymptote at y = n , where a n and b m
bm
are the leading terms of p(x) and q(x), respectively, if
p(x) and q(x) have the same degree;
(c) no horizontal asymptote, if the degree of p(x) is
greater than the degree of q(x).
f (x)
is of the
x→c g(x)
87. Prove the second part of Theorem 1.29: If lim
form
1
f (x)
, then lim
= −∞.
x→c g(x)
0−
x→∞
x→∞
lim ( f (x) + g(x)) = L + M.
x→∞
90. Prove the first part of Theorem 1.31(a): If k > 0, then
lim x k = ∞. (Hint: Given M > 0, choose N = M1/k . Then
x→∞
show that for x > N it must follow that x k > M.)
91. Prove the second part of Theorem 1.31(a): If k > 0, then
lim x−k = 0.
x→∞
92. Prove the k = 1 case of the first part of Theorem 1.31(b):
that lim e x = ∞. (Hint: Given M > 0, choose N = ln M.
x→∞
Then if x > N = ln M, we must have x = ln M + c for some
positive number c. Use this to show that e x > M.)
93. Prove the k = 1 case of the second part of Theorem 1.31(b): that lim e −x = 0.
x→∞
1 − cosθ
94. Prove that lim
= 0 by using the double-angle
θ
θ →0
identity cos 2θ = 1 − 2 sin2 θ and the other special
f (x)
is of
x→∞ g(x)
88. Prove the second part of Theorem 1.30: If lim
the form
89. Prove that the sum rule for limits also applies for
limits as x → ∞: If lim f (x) = L and lim g(x) = M, then
trigonometric limit lim
θ →0
sinθ
= 1.
θ
1
f (x)
, then lim
= 0.
x→∞ g(x)
−∞
Thinking Forward
A limit representing an instantaneous rate of change: After
t seconds, a bowling ball dropped from 350 feet has height
h(t) = 350 − 16t 2 , measured in feet.
In the following exercise you will investigate the convergence
of this limit and also get a preview of Taylor series, which we
will see in Chapter 8.
Use the substitution n =
1
to show that the preceding
h
limit statement is equivalent to the limit statement
n
1
lim 1 +
= e.
n→∞
350 ft
n
Calculate the average rate of change of the height of
the bowling ball from t = 3 to t = 3 + h seconds in
the cases where h is equal to 0.5, 0.25, 0.1, and 0.01.
The Binomial Theorem says that an expression of the
form (a + b) n can be expanded to
n n 0
n n−1 1
n n−1 2
n 0 n
a b +
a b +
a b + ··· +
x y ,
0
1
2
n
where, for any 0 ≤ k ≤ n, the symbol kn is equal to
Write down a formula for the average rate of change
of the height of the bowling ball from time t = 3 to
time t = 3 + h, assuming that h > 0. The only letter
in your formula should be h.
tegers from 1 to n. By convention we set 0! = 1. Apply
1 n
this expansion to the expression 1 +
.
h
Take the limit as h → 0+ of the formula you found for
average rate of change in the previous problem. What
does this limit represent in real–world terms?
n!
. Here n! is n factorial, the product of the ink!(n − k)!
1+
Taylor Series: In this section we learned that e can be thought
of as the following limit:
lim (1 + h)
h→0
1/ h
= e.
n
Show that as n → ∞ we would expect the preceding
expansion to approach
1
1
1
1
1
+ + + + + ··· .
1!
2!
3!
4!
5!
(Hint: Think about limits of rational functions and ratios
of leading coefficients.)
Use a calculator to find the sum of the first six terms
of the sum from the previous problem, and compare
this sum with your calculator’s best approximation of
the number e.
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
152
Chapter 1
November 21, 2012
Limits
CHAPTER REVIEW, SELF-TEST, AND CAPSTONES
Before you progress to the next chapter, be sure you are familiar with the definitions, concepts, and basic skills outlined here.
The capstone exercises at the end bring together ideas from this chapter and look forward to future chapters.
Definitions
Give precise mathematical definitions or descriptions of each
of the concepts that follow. Then illustrate the definition with
a graph or algebraic example, if possible.
the intuitive meaning of the limit statements
lim f (x) = L, lim− f (x) = L, and lim+ f (x) = L
x→c
x→c
x→c
the intuitive meaning of the
lim f (x) = ∞ and lim f (x) = L
x→c
x→c
the formal δ–M, N–, and N–M definitions of the
limit statements lim f (x) = ∞, lim f (x) = L, and
x→c
x→∞
x→c
statements
the formal δ– definition of the limit statements
lim f (x) = L, lim− f (x) = L, and lim+ f (x) = L
x→c
limit
x→∞
what it means, in terms of limits, for a function f to be continuous at a point x = c, left continuous at x = c, and right
continuous at x = c
what it means for a function f to be continuous on a closed
interval [a, b], or continuous on an open interval (a, b), or
continuous on a half-closed interval [a, b)
what it means, in terms of limits, for a function to
have a removable discontinuity, a jump discontinuity, or an
infinite discontinuity at x = c
the definition of the number e in terms of a limit
Power functions are continuous everywhere, which
.
means in terms of limits that
All algebraic functions are
on their domains, which
means in terms of limits that if x = c is in the domain of
.
an algebraic function f , then
All basic transcendental functions are
on their domains, which means in terms of limits that if x = c is in
the domain of a basic exponential, logarithmic, trigono.
metric, or inverse trigonometric function f , then
If lim g(x) exists and f (x) is a function that is
x→∞
lim f (x) = ∞, respectively
what we mean when we say that a limit exists, or that a
limit does not exist
what it means, in terms of limits, for a function f to have
a vertical asymptote at x = c or a horizontal asymptote at
y=L
Theorems
Fill in the blanks to complete each of the following theorem
statements:
If lim f (x) = L and lim f (x) = M, then
x→c
x→c
lim f (x) = L if and only if lim− f (x) =
x→c
lim+ f (x) =
x→c
x→c
.
and
.
For δ > 0, x ∈ (c − δ, c) ∪ (c, c + δ) if and only if
< δ.
0<
For > 0, f (x) ∈ (L − , L + ) if and only if
The Extreme Value Theorem: If f is
on a closed interval [a, b], then there exist values M and m in the interval
and f (m) is
.
[a, b] such that f (M) is
< .
The Intermediate Value Theorem: If f is
on a closed
and
interval [a, b], then for any K strictly between
, there exists at least one c ∈ (a, b) such that
.
A function f can change sign from positive to negative, or
,
, or
at
vice versa, at x = c only if f (x) is
x = c.
Constant, identity, and linear functions are continuous
,
everywhere, which means in terms of limits that
, and
.
x→c
to g(x) for all x sufficiently close to
, but not
, then
.
necessarily at
The Squeeze Theorem for Limits: If l(x) ≤ f (x) ≤ u(x) for
, but not necessarily
all x sufficiently close to
, and if lim l(x) and lim u(x) are both equal
at
x→c
to L, then
x→c
.
p(x)
is a rational function with deg(p(x)) = n
q(x)
p(x)
and deg(q(x)) = m. If n < m, then lim
=
;
x→∞ q(x)
p(x)
if n = m, then lim
=
; and if n > m, then
x→∞ q(x)
p(x)
lim
=
.
x→∞ q(x)
Suppose
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
December 1, 2012
153
Chapter Review, Self-Test, and Capstones
Limit Rules and Indeterminate Forms
Limits of basic functions: Fill in the blanks to complete the limit
rules that follow. You may assume that k is positive.
lim k =
lim x =
lim(mx + b) =
lim Ax =
x→c
x→c
lim x k =
x→∞
x→c
lim( f (x) + g(x)) =
n
x→c
x→c
x→c
lim kf (x) =
x→c
x→c
Limits of combinations: Fill in the blanks to complete the limit
rules that follow. You may assume that k and c are any real
numbers and that both lim f (x) and lim g(x) exist.
lim( f (x) − g(x)) =
x→c
lim x−k =
lim( f (x)g(x)) =
x→∞
x→c
lim b =
x
x→c
lim sin x =
x→c
lim
x→c
lim ekx =
lim 2 x =
lim ln x =
x→∞
x→∞
x→0+
lim tan−1 x =
x→∞
lim f (g(x)) =
lim (0.75) x =
Indeterminate forms: Identify which of the limit forms listed
here are indeterminate. For each form that is not indeterminate, describe the behavior of a limit of that form.
x→∞
1
0+
1
0−
1
∞
1
−∞
0
1
0
0
∞
1
0
∞
∞
∞
∞
0+
0·∞
∞·∞
∞+∞
∞−∞
00
∞
∞
∞0
−∞
∞−∞
lim tan−1 x =
x→−∞
lim (1 + x)1/x =
lim
sin x
=
x
lim
x→0
x→0
1 − cos x
=
x
, provided that
x→c
lim ln x =
x→∞
ex − 1
=
x
x→0
, provided that
lim e−kx =
x→∞
lim
x→0
f (x)
=
g(x)
∞(−∞)
0
01
∞
∞
∞
1
1
0
Skill Certification: Basic Limits
Calculating limits: Find each limit by hand.
1. lim 3x−4
x→0
3. lim
x→2
5.
7.
9.
11.
13.
1
2−x
2x 3 − x 2 − 2x + 1
lim
x→1
x 2 − 2x + 1
3x − 4x
lim
x→0
3x
ln x
lim
x→0+ x
1 − ex
lim
x→∞
e2x
√
lim ( x − x)
x→∞
15. lim
x→3
1
1
−
x−3
x
x−3
2.
17.
lim −2x−1/2
x→∞
4. lim
x→1
1
x2 − 1
x 3 + 2x − 1
6. lim
x→−∞
1 − x4
ex − 1
8. lim 2x
x→0 3e − 2e x − 1
x−1
10. lim ln
2
x→∞
1 − 3x
sin x
x
√
x − x3
14. lim+
x→0
x2
√
2− x
16. lim
x→4 4 − x
12.
lim
x→π/2
lim (−2x 3 + x 2 − 10)
x→∞
27.
(2x − 1)(x + 1)
x→∞
x2 − 4
√
x
lim
√
x→∞ 1 −
x
3x
lim
x→0 sin 2x
1 − cos x
lim
x→0
sin x
x
1
lim 1 +
29.
lim sin x
18. lim(x−3 − 2x−1 )
x→0
2
19.
21.
23.
25.
31.
lim
x
x→∞
x→∞
lim
x→∞
1
sin x
x
20.
22.
lim
x→∞
lim e x tan−1 x
x→−∞
24. lim
x→0
26.
(x − 1)(3x + 1)3
(x − 2)4
sin2 3x
x
lim sin(tan−1 x)
x→∞
x2
−1
1
30. lim sin
x→∞
x
1
32. lim x sin
x→0
x
28. lim
x→0 e x
CONFIRMING PAGES
12:7
TKmaster2010
WHF00153/FREE087-Taalman
154
Chapter 1
November 21, 2012
Limits
Capstone Problems
A.
B.
C.
Continuity of piecewise-defined functions: For each given
function f , find a real number a that makes f continuous
at x = 0, if possible.
3x + 1, if x < 0
(a) f (x) =
2x + a, if x ≥ 0
⎧ a
⎪
⎨ x + 2 , if x < 0
(b) f (x) =
3, if x = 0
⎪
⎩
ax + 1, if x > 0
Limits that define derivatives: In the next chapter we will be
interested in derivatives, which we will define as limits
of the form
f (c + h) − f (c)
.
lim
h→0
h
(a) Calculate this limit for f (x) = x 3 and c = 0.
(b) Calculate this limit for f (x) = x 3 and c = 2.
(c) Calculate this limit for f (x) = x 3 and general c = x.
This time your answer will be a function of x instead
of a number.
The limit of a model at infinity: Leila is interested in the
effect of a stabilized wolf population on the eventual
population of beavers in Idaho. The following table gives
estimated beaver populations B(t) for t = 0, 1, 2, 3, 4, and
5 years after 2005:
t
0
1
2
3
4
5
B(t) 48,112 42,256 47,088 43,684 46,320 44,704
M(t) = −51x 2 + 918x + 41, 389
is a good model for the relative maximum data points
at t = 0, 2, and 4, and that
m(t) = 33.25x 2 − 583.5x + 48, 122
is a good model for the relative minimum data points
at t = 1, 3, and 5. Verify that these functions do in fact
pass through the relevant data points, and graph the
data for B(t) along with the two functions.
(b) Do the two quadratics M(t) and m(t) ever meet? If
so, where? What conclusion could Leila make concerning the eventual steady population lim B(t) of
t→∞
beavers in Idaho?
D. The limit of a rational function model at infinity: Upon further reflection, Leila decides that the quadratics used
in the previous problem are unreasonable, since the
quadratic model for the relative maximum values could
be interpreted as indicating that the eventual number of
beavers would be unbounded. She decides to change her
model for the relative maximum beaver populations to
M(t) =
40944 t 2 + 454512 t − 1732032
.
t 2 + 9 t − 36
(a) Verify that this function does pass through the data
points at t = 0, 2, and 4. Is this function continuous
everywhere? (Hint: Consider lim M(t).)
t→3
(b) Compute lim M(t). What is the significance of this
t→∞
(a) Leila makes a plot of these values of B(t) and notes
that the population of beavers is cyclical with diminishing amplitude. She finds that the quadratic
function
number?
CONFIRMING PAGES
15:42
TKmaster2010
WHF00153/FREE087-Taalman
December 1, 2012
C H A P T E R 2
Derivatives
2.1
An Intuitive Introduction to Derivatives
Slope Functions
Position and Velocity
Approximating the Slope of a Tangent Line
Approximating an Instantaneous Rate of Change
Examples and Explorations
2.2
Formal Definition of the Derivative
The Derivative at a Point
The Derivative as a Function
Differentiability
Tangent Lines and Local Linearity
Leibniz Notation and Differentials
Examples and Explorations
2.3
f (x + h) − f (x)
h→0
h
lim
Rules for Calculating Basic Derivatives
Derivatives of Linear Functions
The Power Rule
The Constant Multiple and Sum Rules
The Product and Quotient Rules
Examples and Explorations
2.4
d k
(x ) = kx k−1
dx
The Chain Rule and Implicit Differentiation
df
df du
=
dx
du dx
Differentiating Compositions of Functions
Implicit Differentiation
Examples and Explorations
2.5
Derivatives of Exponential and Logarithmic
Functions
Derivatives of Exponential Functions
Exponential Functions Grow Proportionally to Themselves
Derivatives of Logarithmic Functions
Derivatives of Inverse Functions*
Examples and Explorations
2.6
Derivatives of Trigonometric and Hyperbolic
Functions
Derivatives of Trigonometric Functions
Derivatives of Inverse Trigonometric Functions
Hyperbolic Functions and Their Derivatives*
Inverse Hyperbolic Functions and Their Derivatives*
Examples and Explorations
Chapter Review, Self-Test, and Capstones
CONFIRMING PAGES
13:41
WHF00153/FREE087-Taalman
156
2.1
Chapter 2
November 21, 2012
Derivatives
AN INTUITIVE INTRODUCTION TO DERIVATIVES
Associated slope functions, tangent lines, and secant lines
Velocity and other instantaneous rates of change
Approximating slopes of tangent lines and instantaneous rates of change
Slope Functions
We begin our study of the derivative with an intuitive introduction in terms of slopes and
rates of change. We also start thinking about how one might calculate, or at least approximate, derivatives. In Section 2.2 we will give a formal mathematical definition of the
derivative in terms of limits.
Intuitively speaking, if the graph of a function f is smooth on an interval (a, b)—
meaning that it does not have any corners, cusps, jumps, or holes—then at every point
(x, f (x)) on the the graph of f on the interval (a, b) we can consider the direction, or
slope, of the function at that point. For example, in the figure that follows at the left, the
tangent line drawn at x = −1 points in the same direction as the function f at the point
(−1, f (−1)). More precisely, if you imagine yourself in a tiny car driving along the graph of
f with your headlights on, then that line represents the direction that your headlights are
pointing when you reach the point (−1, f (−1)) from the right or the left. Similarly, the line
drawn at x = 4 represents the direction of the graph of f at the point (4, f (4)).
Tangent lines at x = −1 and x = 4
Heights of y = f (x)
Slopes of y = f (x)
y
y
y
slope 0
8
x
3
1
1
3
5
e5
5
slop
3
4
1
5
height 8
e
1
8
slop
3
slope
TKmaster2010
height 0
x
4 2
height 0
2
4
slope 0
height 5
4
6
x
height 4
We can use the graph of a smooth function f to define a new function whose output
at each point x is the slope of the tangent line at x. For example, the slopes shown on the
graph of f in the middle figure are used to define heights on the graph of the associated
slope function shown at the right. This associated slope function is what we will define
in Section 2.2 as the derivative of f (x) and denote as f (x) (pronounced “f prime of x”).
Notice the following relationships between a function f and its derivative f :
For each x, the slope of f (x) is the height of f (x).
Where f has a horizontal tangent line, the derivative f has a root.
Where the graph of f is increasing, the derivative f is above the x-axis.
Where the graph of f is decreasing, the derivative f is below the x-axis.
Where f has steep slope, the derivative f has large magnitude.
Where f has shallow slope, the derivative f has small magnitude.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.1
157
An Intuitive Introduction to Derivatives
Position and Velocity
Suppose an object moves in a straight path so that after t seconds it is a distance of s(t)
units from its starting point. We will call the function s(t) describing the motion of the
object a position function. The moving object has a speed and a direction at any time t,
and the combination of these two measurements defines a velocity function for the object.
Specifically, if we consider one direction on the straight path as the “positive” direction and
the other as the “negative” direction, then the velocity of the object at time t is the speed of
the object times either +1 or −1, depending on the direction in which the object is moving.
The velocity v(t) of such a moving object is a measurement of how the position function of the object is changing over time. Intuitively, because the way position changes at
a particular moment in time is measured by the slope of its graph, velocity is the associated slope function for position. In other words, velocity is the derivative of position, or in
symbols, v(t) = s (t). Similarly, acceleration a(t) measures how velocity changes, and thus
a(t) = v (t). We will examine these relationships more precisely in Section 2.2.
For example, suppose you throw a grapefruit straight up into the air, releasing it at a
height of 4 feet and an upwards velocity of 32 feet per second, as illustrated in the figure
that follows on the left. On the right is a plot of the height of the grapefruit over time. The
points A, B, C, D, and E show the height s(t), in feet, of the grapefruit at t = 0, t = 0.6, t = 1,
t = 1.75, and t = 2.118 seconds. At A, the grapefruit is moving upwards quickly. Because
of the downwards pull of gravity, the grapefruit is moving upwards more slowly at B. At C
the grapefruit is at the top of its flight and about to fall to the ground. Gravity then causes
the grapefruit to fall faster and faster through D and then finally E when it hits the ground.
Position increases, then decreases
s
C
C
20
B
B
D
D
32 ft/s
A
4
4 ft
E
A
E
0.6
1
1.75 2.118
t
The next leftmost figure shows the velocity v(t) of the grapefruit. Notice that at times A
and B, when the grapefruit is moving upwards, its velocity is positive; at C, when the
grapefruit turns around at the top of its flight, its velocity is zero; and at D and E, when
the grapefruit is falling to the ground, its velocity is negative. The rightmost figure shows
the constant acceleration of the grapefruit due to gravity.
Velocity is positive, then negative
Acceleration is constant
a
v
32
A
B
C
0.6
1
1.75 2.118
D
35.8
E
0.6
1
1.75 2.118
B
C
D E
t
32
A
CONFIRMING PAGES
t
22:0
TKmaster2010
WHF00153/FREE087-Taalman
158
Chapter 2
November 21, 2012
Derivatives
In general, our earlier list of relationships between a function f (x) and its derivative
f (x) translates into a list of relationships between position s(t) and velocity v(t) = s (t) as
follows:
For each t, the way position s(t) is changing is measured by velocity v(t).
When position s(t) is not changing, velocity v(t) is zero.
When position s(t) is increasing, velocity v(t) is positive.
When position s(t) is decreasing, velocity v(t) is negative.
When position s(t) is changing rapidly, velocity v(t) has large magnitude.
When position s(t) is changing slowly, velocity v(t) has small magnitude.
Approximating the Slope of a Tangent Line
Usually, calculating the slope of a line is a simple matter: Simply take two points (x0 , y0 )
and (x1 , y1 ) on the line and calculate the “rise over run,” which is equal to the average
y
y −y
rate of change
= 1 0 . With tangent lines the situation is more complicated, because
x
x1 −x0
we know only one point on a tangent line, namely, the point (c, f (c)) where it touches the
function. The slope of the tangent line measures the “direction” of the function, but how do
we calculate that from only one point? The key will be to use nearby points on the function
to approximate nearby slopes.
The secant line from a to b for a function f is the line that passes through the points
(a, f (a)) and (b, f (b)). If f is a smooth function and x = z is a point that is close to x = c, then
the slope of the secant line from x = c to x = z will be close to the slope of the tangent line
to f at x = c, as shown in the middle graph that follows:
Secant line from
(c, f (c)) to (z, f (z))
Tangent line at (c, f (c))
y
y
6
y
6
5
f(c)
Secant line from
(c, f (c)) to (c+h, f (c+h))
f (z)
f (c)
4
6
tangent
5
secant
4
f (c h)
f (c)
3
3
2
2
2
1
1
1
2
3
4
5
6
c
1
1
2
3
c
z
4
5
6
x
secant
4
3
x
tangent
5
h
1
2
3
4
5
6
x
c ch
If we choose points z that are closer and closer to the point x = c, we will get secant lines
that get closer and closer to the tangent line we are interested in. Equivalently, we could
think of the second point z as “c plus a little bit,” where the little bit is called h. In other
words, z = c + h, as in the rightmost graph shown.
Since we know two points on a secant line, we can easily calculate its slope. The slope
f (c) of the tangent line to f at x = c can be approximated by the slope of a nearby secant
line from x = c to x = z, or equivalently, from x = c to x = c + h:
f (c) ≈
f (z) − f (c)
,
z−c
or equivalently,
f (c) ≈
f (c + h) − f (c)
.
h
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.1
An Intuitive Introduction to Derivatives
159
The preceding expressions are often called difference quotients, and when we find the
derivative of a function f , we say that we are differentiating the function. If the graph of f
is smooth, then as z gets closer to c, or as h gets closer to 0, these approximations get closer
and closer to f (c). In Section 2.2 we will in fact take the limit as z → c, or, equivalently, as
h → 0, to define the derivative exactly.
Approximating an Instantaneous Rate of Change
You may have noticed that the approximations we have been using for f (c) are closely
related to the formula for average rate of change. This is no coincidence, since average rates
of change are in fact the same as slopes of secant lines. As we choose points z = c + h closer
and closer to x = c, these average rates of change approach the instantaneous rate of
change of the function at x = c. For a general function f , this instantaneous rate of change
at x = c is the derivative, that is, the slope f (c) of the tangent line to f at x = c. In the case
of a position function s(t), the instantaneous rate of change is the velocity v(c) = s (c).
We can approximate instantaneous rates of change in a position or velocity context in
much the same way as we approximated slopes of tangent lines in the previous discussion.
Suppose an object is moving along a straight path. The distance formula says that for
such a moving object, the distance travelled, average rate, and time elapsed are related by
the formula d = rt (“distance equals rate times time”). We can also write this formula as
d
d
r = , or more accurately, as r = , since we want to consider the change in distance over
t
t
a corresponding change in time. If an object starts at position s0 at time t0 and ends at
position s1 at time t1 , then we have
s1 − s 0
average velocity average rate of change of d
=
=
.
=
position from t0 to t1
from t0 to t1
t
t1 − t 0
Now suppose s(t) describes the position of the object at time t and we are interested in finding the velocity v(c) at some time t = c. This instantaneous velocity can be approximated
by the average velocity over a small time interval [c, z], or equivalently, [c, c + h]:
v(c) ≈
s(z) − s(c)
,
z−c
or equivalently,
v(c) ≈
s(c + h) − s(c)
.
h
Notice that this is just a special case of what we did earlier for a general function f (x) and
its derivative f (x) at a point x = c.
We can use derivatives to examine instantaneous rates of change in many contexts. In
general, the derivative of a function y(x) represents the instantaneous rate of change of the
variable y as the variable x varies. The units for the derivative y (x) are the units for the
variable y divided by the units for the variable x. For example, if time t is measured in hours
and position s(t) is measured in miles, then the velocity v(t) = s (t) is measured in miles
per hour. As another example, if Q(t) is the amount of money in a savings account after
t years, measured in dollars, then Q (t) is the rate at which the savings balance changes
over time, with units measured in dollars per year.
Examples and Explorations
EXAMPLE 1
Sketching the graph of an associated slope function
Given the following graph of the smooth function f , sketch the graph of its associated slope
function f :
CONFIRMING PAGES
22:0
WHF00153/FREE087-Taalman
160
Chapter 2
November 21, 2012
Derivatives
y
3
2
1
2
1
1
1
2
3
x
4
2
3
SOLUTION
A good place to begin is by marking all the locations on the graph of f where the tangent
line is horizontal and thus has slope 0. In this case that happens at x = 0 and at x = 2, as
shown next at the left. Thus the associated slope function f has zeroes at x = 0 and x = 2,
as shown next at the right.
f (x) heights are f (x) slopes
f (x) with slopes marked
y
y
3
3
2
2
zero
po
.
neg
.
neg
1
s.
2
1 zero
1
1
2
3
4
pos.
1
zero
x
2
1
2
zero
1
1
2
3
4
x
2
neg.
3
neg.
3
Looking again at the graph of f , we see that its tangent lines have positive slope between
x = 0 and x = 2 (see, for example, the positive slope marked at x = 1). This means that,
in the graph of f , the heights will be positive between x = 0 and x = 2. Similarly, the
negative slopes on the graph of f to the left of x = 0 and to the right of x = 2 correspond
to negative heights on the graph of f .
EXAMPLE 2
Graphing velocity from the graph of position
Suppose the graph that follows describes your distance from home one morning as you
drive back and forth from your sister’s house. Describe a possible scenario for your travels
that morning. Then sketch the corresponding graph of your velocity.
Distance from home
TKmaster2010
a
b
c
d
e
f
g
Time
CONFIRMING PAGES
22:0
WHF00153/FREE087-Taalman
November 21, 2012
2.1
161
An Intuitive Introduction to Derivatives
SOLUTION
One possible scenario is this: You drive to your sister’s house for a visit. After talking to
her for a few minutes, you realize you forgot something at home and race back to get
it. In the middle of returning to your sister’s house, you have to stop at a red light for a
couple of minutes. Following the times marked on the time axis, we see that from 0 to a
you drive to your sister’s house, you talk until b, race home from b to c, leave your house
at d and get stopped at the light at e, and move on at f until you get back to your sister’s
house at g.
Distance from home
a
b
c
d
e
f
g
Time
The graph of your velocity that morning is the graph of the associated slope function
for the given position graph. The slope of the position graph is positive from 0 to a, zero
from a to b, negative and steep from b to c, zero from c to d, positive and steep from d to e,
zero from e to f , positive and steep from f to g, and finally zero again after g. The previous
sentence also describes the height of the corresponding velocity graph, where steep slope
values correspond to large magnitudes of velocity:
Velocity
TKmaster2010
a
b
c
d
e
f
g
Time
EXAMPLE 3
Estimating the slope of a tangent line with a sequence of secant lines
1
Estimate the slope of the line tangent to the graph of f (x) = − x 2 + 3x at the point (2, f (2))
2
by calculating a sequence of slopes of secant lines.
SOLUTION
1
The tangent line passes through the point (2, f (2)) = 2, − (2)2 + 3(2) = (2, 4) and is
2
shown in red in each of the graphs that follow. Since we only know one point on this
line, we cannot compute its slope directly. We will approximate its slope by considering a
sequence the slopes of secant lines on smaller and smaller intervals, namely, [2, 3], [2, 2.5],
[2, 2.25], and [2, 2.1], as shown in the following four graphs below:
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
162
Chapter 2
November 21, 2012
Derivatives
Secant line on [2, 3]
Secant line on [2, 2.25]
Secant line on [2, 2.5]
y
y
Secant line on [2, 2.1]
y
y
6
6
6
6
5
5
5
5
4
4
4
4
3
3
3
3
2
2
2
2
h
1
1
h
1
2
3
x
z
4
x
1
2
h
1
3
4
x
1
x z
2
h
1
3
4
x
1
2
xz
3
4
x
xz
In our difference quotient notation, these intervals correspond to a sequence of points
z = 3, z = 2.5, z = 2.25, and z = 2.1 that approach c = 2. Equivalently, we can think of this
sequence as a series of h-values h = 1, h = 0.5, h = 0.25, and h = 0.1 approaching zero.
The slope of the secant line from x = 2 to x = 3 in the leftmost graph is given by the
difference quotient:
1
1
− (3)2 + 3(3) − − (2)2 + 3(2)
f (3) − f (2)
4.5 − 4
2
2
=
=
= 0.5.
3−2
3−2
3−2
Similarly, the slopes of the remaining three secant lines are given by the difference quotients:
f (2.5) − f (2)
= 0.75,
2.5 − 2
f (2.25) − f (2)
= 0.875,
2.25 − 2
and
f (2.1) − f (2)
= 0.95.
2.1 − 2
Each of these slopes is an approximation to the slope of the red tangent line. As the graphs
shown suggest, we would expect this sequence of slopes to be getting closer and closer to
the actual slope of the red tangent line; notice for example that, in the last figure shown,
the green secant line is almost indistinguishable from the red tangent line.
In a similar fashion we can calculate the slopes of secant lines from the left of x = 2. For
example, the slope of the secant line from x = 1 to x = 2 is given by the difference quotient:
1
1
− (1)2 + 3(1) − − (2)2 + 3(2)
f (1) − f (2)
2.5 − 4
2
2
=
=
= 1.5.
1−2
1−2
1−2
Over the smaller intervals [1.5, 2], [1.75, 2], and [1.9, 2] we have secant lines with slopes
given by
f (1.5) − f (2)
= 1.25,
1.5 − 2
f (1.75) − f (2)
= 1.125,
1.7 − 2
and
f (1.9) − f (2)
= 1.05.
1.9 − 2
Putting all this information together, we obtain the following table:
Interval
Slope
[1, 2]
[1.5, 2]
[1.75, 2]
[1.9, 2]
*
[2, 2.1]
[2, 2.25]
[2, 2.5]
[2, 3]
1.5
1.25
1.125
1.05
*
0.95
0.875
0.75
0.5
From this table, we might guess that the slope of the tangent line is 1. This guess is only
an estimate; the slope of the tangent line might instead be something like 0.97 or 1.02, but
we don’t have enough information to say otherwise at this point.
EXAMPLE 4
Estimating instantaneous velocity with a sequence of average velocities
It can be shown that a watermelon dropped from a height of 100 feet will be s(t) =
− 16t 2 + 100 feet off the ground t seconds after it is dropped. Approximate the instantaneous velocity of the watermelon at time t = 1 by calculating a sequence of average velocities. Then interpret these average velocities graphically as slopes of secant lines.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.1
163
An Intuitive Introduction to Derivatives
SOLUTION
To estimate the instantaneous velocity at t = 1 we will look at a sequence of small time
intervals near t = 1 and consider the corresponding average velocities. The time intervals
we choose to consider are [1, 2], [1, 1.5], [1, 1.25], and [1, 1.1]. These intervals correspond
to z = 2, 1.5, 1.25, and 1.1, or equivalently, to h = 1, 0.5, 0.25, and 0.1.
Let’s look first at the interval [1, 2]. At t = 1 the watermelon is s(1) = 84 feet from the
ground, and at t = 2 the watermelon is s(2) = 36 feet from the ground, as illustrated
here:
100
84
t 1 to t 2
36
0
The average velocity over the first interval is therefore given by the difference quotient:
s(2) − s(1)
(−16(2)2 + 100) − (−16(1)2 + 100)
=
= 36 − 84
2−1
1
= −48 feet per second.
Similarly, the average velocities over the remaining three time intervals, in feet per second,
are
s(1.5) − s(1)
= −40,
1.5 − 1
s(1.25) − s(1)
= −36,
1.25 − 1
s(1.1) − s(1)
= −33.6.
1.1 − 1
and
Each of these average velocities is an approximation to the instantaneous velocity of the
watermelon at time t = 1. Since the approximations should be improving as z gets closer
to 1 (or, equivalently, as h gets closer to 0), we might guess that the instantaneous velocity
of the watermelon at time t = 1 is some value greater than, but close to, −33.6 feet per
second. For example, we might estimate that the instantaneous velocity at time t = 1 is
−33 feet per second.
Each average velocity just calculated is an average rate of change of position, and thus
can be thought of as the slope of a secant line, as in the four graphs shown next. As we
consider smaller and smaller time intervals, we see that the slopes corresponding to these
average velocities approach the slope of the red tangent line to s(t) at t = 1, which in turn
represents the instantaneous velocity at t = 1.
Average rate of change
of s(t) on [1, 2]
Average rate of change
of s(t) on [1, 1.5]
y
Average rate of change
of s(t) on [1, 1.25]
y
Average rate of change
of s(t) on [1, 1.1]
y
y
100
100
100
100
75
75
75
75
50
50
50
50
25
25
1
2
t
th
3
x
25
1
2
t th
3
x
25
1
2
3
x
t th
CONFIRMING PAGES
1
2
tt h
3
x
22:0
TKmaster2010
WHF00153/FREE087-Taalman
164
Chapter 2
November 21, 2012
Derivatives
Why does it make intuitive sense that when the graph of a smooth function f has a
TEST YOUR
? UNDERSTANDING
horizontal tangent line, the graph of its associated slope function will have a root?
Why does it make intuitive sense that when the graph of a smooth function f is
increasing, the graph of its associated slope function will be above the x-axis?
In our initial discussion of position and velocity in this section, we showed graphs for
the position, velocity, and acceleration of a grapefruit thrown into the air. Why does it
make sense that the graph of acceleration is constant, given that the graph of velocity
is a straight line?
How can we use a sequence of slopes of secant lines to estimate the slope of a tangent
line? Why is considering a sequence of secant lines better than considering just one
secant line?
In real–world examples, how are the units of a derivative of a function related to the
units of the independent and dependent variables of that function?
EXERCISES 2.1
Thinking Back
Slope and linear functions: If f is a linear function with
slope −3 such that f (2) = 1, find the following, without
first finding an equation for f (x).
• f (4)
• f (7)
• f (−2)
Approximating limits: Use sequences of approximations
to estimate the values of
4 − x2
z 3 − 27
• lim
• lim
x→2 2 − x
z→3 z − 3
Identifying increasing and decreasing behavior: Use a
graphing utility to determine the intervals on which
f (x) = −4x 5 + 25x 4 − 40x 3 is increasing or decreasing.
Interpreting distance graphically: When flying home for
the holidays, Eva often flies between Denver International Airport (DIA) and Chicago O’Hare (ORD).
Suppose Eva’s plane takes off from DIA and 50 miles
from ORD the plane has to circle the airport because
of snow. The plane circles ORD four times and then
lands.
(a) Draw a graph depicting the distance from DIA to
Eva’s plane as a function of time.
(b) Draw a graph depicting the distance from ORD to
Eva’s plane as a function of time.
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The slope of the tangent line to a function f at the point x = 4 is given by f (4).
(b) True or False: The instantaneous rate of change of a
function f at the point x = −3 is given by f (−3).
(c) True or False: The instantaneous rate of change of a
function f at a point x = a can be represented as the
slope of a secant line.
(d) True or False: Where a function f is positive, its
associated slope function f is increasing.
(e) True or False: Where a function f is decreasing, its
associated slope function f is negative.
(f) True or False: When a function f has a steep slope at
a point on its graph, its instantaneous rate of change
at that point will have a large magnitude.
(g) True or False: When the graph of a function f is decreasing with a steep slope, the graph of the associated slope function f is negative with a large
magnitude.
(h) True or False: Suppose an object is moving in a straight
path with position function s(t). If s(t) is positive and
decreasing, then the velocity v(t) is negative.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function whose associated slope function f is positive on (−∞, 2) and negative on (2, ∞).
(b) The graph of a function with the following three
properties: The average rate of change of f on [0, 2]
is 3, the average rate of change of f on [0, 1] is
−1, and the average rate of change of f on [−2, 2]
is 0.
(c) The graph of a function f with the following three
properties: The instantaneous rate of change of f at
x = 2 is zero, the average rate of change of f on [1, 2]
is 2, and the average rate of change of f on [2, 4]
is 1.
3. Explain why it is not a simple task to calculate the slope of
the tangent line to a function f at a point x = c. Shouldn’t
calculating the slope of a line be really easy? What goes
wrong here?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.1
4. Let l be the line connecting two points (a, f (a)) and
(b, f (b)) on the graph of a function f . What does this line
l have to do with the average rate of change of f on the
interval [a, b], and why?
5. Given that s(t) measures the distance an object has travelled over time, explain what the expression
s(b) − s(a)
has
b−a
to do with the distance formula d = rt.
6. How is velocity different from speed? What does it mean
if velocity is negative?
7. What is the relationship between the derivative of a function f at a point x = c, the slope of the tangent line to the
graph of f at x = c, and the instantaneous rate of change
of f at x = c?
8. On a graph of f (x) = x 2 ,
(a) draw the tangent line to the graph of f at the point
(2, f (2));
(b) draw the secant line from (2, f (2)) to (2.75, f (2.75));
(c) draw the secant line from (1.75, f (1.75)) to (2, f (2)).
(d) Which secant line is a better approximation to the
tangent line, and why?
9. In Example 3 we estimated the slope of the tangent line to
1
2
f (x) = − x 2 + 3x at x = 2. Get a better estimate by calculating the slopes of secant lines with values of z even
closer to x = 2—for example, z = 2.01, z = 2.001, and
z = 2.0001.
10. In Example 3 we estimated the slope of the tangent
1
line to f (x) = − x 2 + 3x at x = 2 by finding slopes of
2
secant lines from x = 2 to various points x = z with z > 2.
Draw a sequence of graphs that illustrates how to do this
for z < 2, and then make specific calculations for z = 1,
z = 1.5, z = 1.75, and z = 1.9. What are the corresponding values of h in this example?
13. Consider again the graph of f at the left. Label each of the
following quantities to illustrate that f (c) ≈
f (c + h) − f (c)
:
h
(c) the slopes
(c) the slopes
g(z) − g(c)
:
z−c
(a) the locations c, z, g(c), and g(z)
(b) the distances z − c and g(z) − g(c)
(c) the slopes
15. For the graph of f shown next at the left, list the following
quantities in order from least to greatest:
(a) the average rate of change of f on [−1, 1]
(b) the instantaneous rate of change of f at x = 1
(c) f (−1)
f (2) − f (−1)
2 − (−1)
(d)
16. For the graph of g(x) shown next at the right, list the
following quantities in order from least to greatest:
(a) the average rate of change of g on [0, 1]
(b) the instantaneous rate of change of g at x = 1
g(−1 + 0.1) − g(−1)
0.1
g(1) − g(−1)
(d)
1 − (−1)
(c)
f (x), Exercises 15 and 17
(a) the locations c, c + h, g(c), and g(c + h)
(b) the distances h and g(c + h) − g(c)
g(c + h) − g(c)
and g (c)
h
f (x), Exercises 11 and 13
g(x), Exercises 12 and 14
y
y
x
x
g(x), Exercises 16 and 18
y
y
2
3
1
2
1
1
2
x
1
2
2
f (c + h) − f (c)
and f (c)
h
g(c + h) − g(c)
:
h
(c) the slopes
g(z) − g(c)
and g (c)
z−c
1
12. For the graph of g(x) appearing next at the right,
label each of the following quantities to illustrate that
g (c) ≈
f (z) − f (c)
and f (c)
z−c
14. Consider again the graph of g(x) at the right. Label
each of the following quantities to illustrate that g (c) ≈
2
(a) the locations c, c + h, f (c), and f (c + h)
(b) the distances h and f (c + h) − f (c)
f (z) − f (c)
:
z−c
(a) the locations c, z, f (c), and f (z)
(b) the distances z − c and f (z) − f (c)
11. For the graph of f appearing next at the left, label each of the following quantities to illustrate that
f (c) ≈
165
An Intuitive Introduction to Derivatives
1
1
2
x
1
17. Consider again the function f graphed at the left. At
which values of x does f have the greatest instantaneous
rate of change? The least? At which values of x is the
instantaneous rate of change of f equal to zero?
18. Consider again the function g(x) graphed at the right. For
which values of x does g(x) have a positive instantaneous
rate of change? Negative? Zero?
19. Make a copy of the graph of f used in Exercises 11 and 13,
and sketch additional secant lines to illustrate that as
h → 0 (or equivalently, as z → c) the slopes of the secant
line get closer and closer to the slope of the tangent line
to f at x = c.
20. The derivative of a smooth function f at a point x = c
can also be approximated with a symmetric difference
quotient:
f (c) ≈
f (c + h) − f (c − h)
.
2h
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
166
Chapter 2
November 21, 2012
Derivatives
(a) Use a graph to illustrate what the symmetric difference measures. Why would it be reasonable to use
the two-sided symmetric difference to approximate
f (c)? (Hint: Your answer should involve a certain kind
of secant line and a discussion of what happens as h gets
close to 0.)
(b) Use a sequence of symmetric difference approximations to estimate the derivative of f (x) = x 2 at x = 3.
Illustrate your answer with a sequence of graphs.
Skills
In Exercises 21–24, sketch the graph of a function f that has
the listed characteristics.
21. f (1) = 2, f (1) = 0, f (3) = 2
22. f (−3) = 0, f (−1) = 0, f (2) = 0
23. f (−1) = 2, f (−1) = 3, f (1) = −2, f (1) = 3
Now go the other way! Each graph in Exercises 31–34
can be thought of as the associated slope function f for
some unknown function f . In each case sketch a possible
graph of f .
y
31.
24. f (−2) = 2, f (0) = 1, f (1) = −5
Sketch a graph of the associated slope function f for each
function f in Exercises 25–30.
y
25.
4
3
3
2
2
1
2
2
1
1
2
3
2
1
1
1
2
x
y
1
1
2
x
2
2
1
1
2
x
3
2
2
1
1
2
x
2
1
1
2
2
3
3
x
For each function f and value x = c in Exercises 35–44,
use a sequence of approximations to estimate f (c). Illustrate
your work with an appropriate sequence of graphs of secant
lines.
35. f (x) = 4 − x 2 , c = 1
36. f (x) = 4 − x 2 , c = 0
2
y
y
30.
2
4
1
3
3
2
2
2
1
1
1
x
2
1
4
2
2
y
3
1
1
1
4
29.
1
x
1
1
34.
1
2
1
1
3
2
2
2
2
y
33.
1
2
x
2
1
28.
2
1
1
x
y
27.
3
1
y
26.
1
y
32.
2
3
2
4
x
4
37. f (x) = x + x 3 , c = 0
38. f (x) = x + x 3 , c = 1
39. f (x) = ln(x 2+1), c = 0
40. f (x) = e x , c = 0
π
41. f (x) = sin x, c =
2
43. f (x) = |x − 1|, c = 3
42. f (x) = arctan x, c = 0
44. f (x) = |x 2 − 4|, c = 1
Applications
A bowling ball dropped from a height of 400 feet will be
s(t) = 400 − 16t 2 feet from the ground after t seconds. Use a
sequence of average velocities to estimate the instantaneous
velocities described in Exercises 45–48.
45. When the bowling ball is first dropped, with h = 0.5,
h = 0.25, and h = 0.1
46. After t = 1 seconds, with h = 0.5, h = 0.25, h = −0.5,
and h = −0.2
47. After t = 2 seconds, with h = 0.1, h = 0.01, h = −0.1,
and h = −0.01
48. When the bowling ball hits the ground, with h = −0.5,
h = −0.2, and h = −0.1
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.1
49. Think about what you did today and how far north
you were from your house or dorm throughout the day.
Sketch a graph that represents your distance north from
your house or dorm over the course of the day, and explain how the graph reflects what you did today. Then
sketch a graph of your velocity.
50. Stuart left his house at noon and walked north on Pine
Street for 20 minutes. At that point he realized he was
late for an appointment at the dentist, whose office was
located south of Stuart’s house on Pine Street; fearing
he would be late, Stuart sprinted south on Pine Street,
past his house, and on to the dentist’s office. When he
got there, he found the office closed for lunch; he was
10 minutes early for his 12:40 appointment. Stuart waited
at the office for 10 minutes and then found out that his
appointment was actually for the next day, so he walked
back to his house. Sketch a graph that describes Stuart’s
position over time. Then sketch a graph that describes
Stuart’s velocity over time.
Dentist
Home
Pine Street
walk
12:30
12:40
12:20
run
walk
51. Every morning Linda takes a thirty-minute jog in Central
Park. Suppose her distance s in feet from the oak tree on
the north side of the park t minutes after she begins her
jog is given by the function s(t) shown that follows at the
left, and suppose she jogs on a straight path leading into
the park from the oak tree.
(a) What was the average rate of change of Linda’s
distance from the oak tree over the entire thirtyminute jog? What does this mean in real-world
terms?
(b) On which ten-minute interval was the average rate
of change of Linda’s distance from the oak tree the
greatest: the first 10 minutes, the second 10 minutes,
or the last 10 minutes?
(c) Use the graph of s(t) to estimate Linda’s average
velocity during the 5-minute interval from t = 5 to
t = 10. What does the sign of this average velocity tell
you in real-world terms?
(d) Approximate the times at which Linda’s (instantaneous) velocity was equal to zero. What is the physical significance of these times?
An Intuitive Introduction to Derivatives
167
(e) Approximate the time intervals during Linda’s jog
that her (instantaneous) velocity was negative. What
does a negative velocity mean in terms of this physical example?
Distance from the oak tree
Distance from the post office
y
y
500
1
400
0.75
300
0.5
200
0.25
100
10
20
30
x
10 20 30 40 50 60
x
52. Last night Phil went jogging along Main Street. His distance from the post office t minutes after 6:00 p.m. is
shown in the preceding graph at the right.
(a) Give a narrative (that matches the graph) of what Phil
did on his jog.
(b) Sketch a graph that represents Phil’s instantaneous
velocity t minutes after 6:00 p.m. Make sure you
label the tick marks on the vertical axis as accurately
as you can.
(c) When was Phil jogging the fastest? The slowest?
When was he the farthest away from the post office?
The closest to the post office?
53. Suppose h(t) represents the average height, in feet, of a
person who is t years old.
(a) In real-world terms, what does h(12) represent and
what are its units? What does h (12) represent, and
what are its units?
(b) Is h(12) positive or negative, and why? Is h (12) positive or negative, and why?
(c) At approximately what value of t would h(t) have a
maximum, and why? At approximately what value of
t would h (t) have a maximum, and why?
54. A tomato plant given x ounces of fertilizer will successfully bear T(x) pounds of tomatoes in a growing
season.
(a) In real-world terms, what does T(5) represent and
what are its units? What does T (5) represent and
what are its units?
(b) A study has shown that this fertilizer encourages tomato production when less than 20 ounces
are used, but inhibits production when more than
20 ounces are used. When is T(x) positive and when
is T(x) negative? When is T (x) positive and when is
T (x) negative?
55. If Katie walked at 3 miles per hour for 20 minutes and
then sprinted at 10 miles an hour for 8 minutes, how fast
would Dave have to walk or run to go the same distance
as Katie did in the same time while moving at a constant
speed? Sketch a graph of Katie’s position over time and
a graph of Dave’s position over time on the same set of
axes.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
168
Chapter 2
November 21, 2012
Derivatives
Katie
t, minutes
a
b
d, distance
Dave
b
d, distance
56. Velocity v(t) is the derivative of position s(t). It is also true
that acceleration a(t) (the rate of change of velocity) is
the derivative of velocity. If a race car’s position in miles
t hours after the start of a race is given by the function
s(t),what are the units of s(1.2)? What are the units and
real-world interpretation of v(1.2)? What are the units
and real-world interpretation of a(1.2)?
57. The total yearly expenditures by public colleges and
universities from 1990 to 2000 can be modeled by the
function E(t) = 123(1.025)t , where expenditures are measured in billions of dollars and time is measured in years
since 1990.
(a) Estimate the total yearly expenditures by these colleges and universities in 1995.
(b) Compute the average rate of change in yearly expenditures between 1990 and 2000.
(c) Compute the average rate of change in yearly expenditures between 1995 and 1996.
(d) Estimate the rate at which yearly expenditures of
public colleges and universities were increasing in
1995.
Proofs
58. Show that if f is a function and z = x + h, then
f (z) − f (x)
f (x + h) − f (x)
=
.
z−x
h
59. Suppose f is a linear function with positive slope. Show
that the average rate of change of f on any interval [a, b]
is positive, and then use this fact to show that f is always
increasing.
Thinking Forward
Taking the limit: We have seen that if f is a smooth function,
then f (c) ≈
f (c + h) − f (c)
. This approximation should get beth
ter as h gets closer to zero. In fact, in the next section we will
define the derivative in terms of such a limit.
f (c + h) − f (c)
.
f (c) = lim
h→0
h
Instead of choosing small values of h, we could have
chosen values of z close to c. What limit involving
z instead of h is equivalent to the one involving h?
Use the limit you just found to calculate the exact
slope of the tangent line to f (x) = x 2 at x = 4. Obviously you should get the same final answer as you did
earlier.
Use the limit just defined to calculate the exact slope
of the tangent line to f (x) = x 2 at x = 4.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
2.2
Formal Definition of the Derivative
169
FORMAL DEFINITION OF THE DERIVATIVE
Using limits to define derivatives, tangent lines, and instantaneous rates of change
Differentiability at a point, from one side, and on intervals
Leibniz notation and multiple derivatives
The Derivative at a Point
In the previous section we examined derivatives intuitively, by discussing tangent lines and
instantaneous rates of change. We will now use limits to make these ideas precise. We have
seen that the slope f (c) of the tangent line to f at x = c can be approximated by the slope
of a nearby secant line from x = c to x = c + h, or equivalently, from x = c to x = z:
f (c) ≈
f (c + h) − f (c)
,
h
or equivalently,
f (c) ≈
f (z) − f (c)
.
z −c
If the limit of these quantities approaches a real number as h → 0, or as z → c, then we
will define that real number to be the derivative of f at the point x = c.
DEFINITION 2.1
The Derivative of a Function at a Point
The derivative at x = c of a function f is the number
f (c) = lim
h→0
f (c + h) − f (c)
,
h
or equivalently,
f (c) = lim
z→c
f (z) − f (c)
,
z −c
provided that this limit exists.
The derivative of a function f at a point x = c measures the instantaneous rate of change of
the function at that point. Notice that this instantaneous rate of change is a limit of average
rates of change.
For example, consider the function f (x) = x 2 . We can calculate the derivative of this
function at the point x = 3 with a limit as h → 0 or with a limit as z → 3. Using the h → 0
definition of the derivative, we have
(3 + h)2 − 32
6h + h2
h(6 + h)
= lim
= lim
= lim (6 + h) = 6.
h
h
h
h→0
h→0
h→0
h→0
f (3) = lim
Using the z → c definition of derivative we obtain the same answer:
f (3) = lim
z 2 − 32
z→c z − 3
= lim
z→c
(z + 3)(z − 3)
= lim(z + 3) = 6.
z−3
z→c
In both calculations above we have shown the instantaneous rate of change of f (x) = x 2 at
x = 3 is equal to 6. At the instant that we have x = 3, the function f (x) = x 2 is changing at
a rate of 6 vertical units for each horizontal unit. This is equivalent to saying that the slope
of the tangent line is equal to 6, as shown here:
Tangent line to f (x) = x 2 at x = 3 has slope f (3) = 6
y
24
21
18
15
12
9
6
3
2 1
1
2
3
4
x
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
170
Chapter 2
November 21, 2012
Derivatives
The Derivative as a Function
By putting all of the point-derivatives of a function together, we can define a function f whose output at any value of x is defined to be the derivative, or instantaneous rate of
change, of f at that point.
DEFINITION 2.2
The Derivative of a Function
The derivative of a function f is the function f defined by
f (x) = lim
The domain of f
f (x + h) − f (x)
,
h
h→0
is the
or equivalently,
f (x) = lim
z→x
f (z) − f (x)
.
z −x
set of values x for which the defining limit of f exists.
The function f is the associated slope function that we investigated in the previous
section, since at each point x its value is the slope of the graph of f . In addition, the function f represents the instantaneous rate of change at every point x. In particular, if s(t) is a
position function, then its instantaneous rate of change is the velocity function v(t) = s (t).
Similarly, the instantaneous rate of change of velocity v(t) is the the acceleration function
a(t) = v (t).
For example, we can calculate the derivative of f (x) = x 2 for all values of x with either
a limit as h → 0 or a limit as z → x; again we choose the first method:
(x + h)2 − x 2
2xh + h2
h(2x + h)
= lim
= lim
= lim (2x + h) = 2x.
h
h
h
h→0
h→0
h→0
h→0
f (x) = lim
Finding f (x) for general x is like calculating f (c) for all possible values x = c at the same
time. Once we have a formula for f (x), we can easily calculate any particular value f (c).
For example, evaluating f (x) = 2x at x = 3 does give us f (3) = 2(3) = 6, as we calculated
before. The following figures show slopes on the graph of f (x) = x 2 together with heights
on the graph of f (x) = 2x, for x = −2, x = 0, and x = 3.
Heights of f (x) = 2x
Slopes of f (x) = x 2
y
y
16
12
12
6
8
4 3 2 1
4
4 3 2 1
4
1
2
3
4
1
2
3
4
x
6
x
12
In this book we will most often use the h → 0 version of the derivative, but will use the
z → x version when it suits our purposes or makes a calculation easier. You will verify that
these two versions of the derivative are equivalent in Exercise 5.
Differentiability
Given a function f and a value x = c, there may or may not be a well-defined tangent line
to the graph of f at (c, f (c)). When there is a well-defined tangent line with a finite slope,
we say that f is differentiable at x = c:
DEFINITION 2.3
Differentiability at a Point
A function f is differentiable at x = c if lim
h→0
f (c + h) − f (c)
exists.
h
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
171
Formal Definition of the Derivative
We used the h → 0 definition of the derivative in Definition 2.3, but everything would work
equally well with the equivalent z → c definition of the derivative. If a point-derivative f (c)
is infinite, then we say that f has a vertical tangent line at x = c. In this case a line exists that
is tangent to the graph of the function, but since that line is vertical, its slope is undefined
and the function fails to be differentiable at that point.
Like continuity, differentiability can be considered from the left or from the right. A
function is left differentiable at x = c if its left derivative exists, and right differentiable
at x = c if its right derivative exists, where the left and right derivatives are defined with
left and right limits as follows:
DEFINITION 2.4
One-sided Differentiability at a Point
The left derivative and right derivative of a function f at a point x = c are, respectively,
equal to the following, if they exist:
f − (c) = lim
h→0−
f (c + h) − f (c)
,
h
f + (c) = lim
h→0+
f (c + h) − f (c)
.
h
We could also use the z → c definition of the derivative to define left and right derivatives,
by considering limits of difference quotients as z → c− and as z → c+ .
For example, consider the function f (x) = |x|. This function has a sharp corner at x = 0,
and therefore we would not expect it to have a well-defined tangent line at that point.
Indeed, when we try to calculate f (0), we encounter the following limit:
|0 + h| − |0|
|h|
= lim .
h
h→0
h→0 h
f (0) = lim
0
0
This limit is initially in the indeterminate form , but we cannot cancel anything in it until
we get rid of the absolute value. Recall that if h ≥ 0, then |h| = h, but if h < 0, then |h| = −h.
Looking from the left and the right, we have the following limits:
f − (0) = lim
|h|
h→0− h
−h
= lim
h→0− h
= lim −1 = −1,
h→0−
|h|
h
f + (0) = lim
= lim = lim 1 = 1.
h→0+ h
h→0+ h
h→0+
Since f − (0) and f + (0) exist but are not equal, f (0) does not exist. The first two graphs
shown next illustrate the left and right derivatives at 0, for small negative h and small
positive h, respectively. The third graph shows the graph of the derivative function f ;
note that this function has a jump discontinuity and is not defined at x = 0.
f has slope −1 as h → 0−
y
y
y
4
4
2
3
3
1
2
2
1
1
sl
e
op
1
3 2 1
f (0) is undefined
f has slope 1 as h → 0+
h
1
2
3
x
3 2 1
pe
3 2 1
1
1
o
sl
h
1
2
3
x
1
2
3
x
2
As with continuity, we say that a function is differentiable on an interval I if it is differentiable at every point in the interior of I, right differentiable at any closed left endpoint,
and left differentiable at any closed right endpoint. For example, the first graph shown
next is differentiable on [2, 3], since it is right differentiable at x = 2, even though it is not
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
172
Chapter 2
November 21, 2012
Derivatives
differentiable at x = 2. The second and third graphs show two more ways that a function
could fail to be differentiable.
Cusp at x = 2
Corner at x = 2
Vertical tangent line at x = 2
y
y
y
4
4
4
3
3
3
2
2
2
1
1
1
1
2
3
4
x
1
2
3
4
x
1
2
3
4
x
In the first and second graphs, the left and right derivatives exist at x = 2 but the two-sided
derivative at x = 2 does not. In the third graph, the left, right, and two-sided derivatives at
x = 2 are all infinite, because the tangent line is vertical.
Another way that a function f can fail to be differentiable at a point x = c is if f fails to
be continuous at x = c. Intuitively, if a function is not continuous, then it has absolutely no
chance of being differentiable; think for a minute about secant lines from the left and the
right at a jump or removable discontinuity, for example. (See also Example 6.) What this
means is that although not every continuous function is differentiable, every differentiable
function is continuous:
THEOREM 2.5
Differentiability Implies Continuity
If f is differentiable at x = c, then f is continuous at x = c.
Proof. If f is differentiable at x = c, then lim
x→c
f (x) − f (c)
= f (c) exists; that is, the limit, and thus f (c),
x−c
is equal to some real number. We will use this fact to show that f is continuous at x = c, by showing
that lim f (x) = f (c). By the sum rule for limits, it is equivalent to show that lim ( f (x) − f (c)) = 0,
x→c
x→c
which we can do by using the expression for the derivative:
f (x) − f (c)
(x − c)
lim ( f (x) − f (c)) = lim
x→c
x→c
x −c
f (x) − f (c)
(lim (x − c))
= lim
x −c
x→c
x→c
←
x −c
= 1 if x = c
x −c
← product rule for limits
= f (c) lim (x − c)
← definition of the derivative
= f (c)(0) = 0.
← limit rules
x→c
Tangent Lines and Local Linearity
Although we have an intuitive sense of the tangent line to a graph at a point, up until now
we did not have a formal mathematical definition for this tangent line. Our geometrical
intuition helped us arrive at an algebraic definition for the derivative, but it is the algebraic
definition that now allows us to define the tangent line precisely. Specifically, we define
the tangent line through a point on the graph of a function to be the line whose slope is
given by the derivative of the function at that point.
THEOREM 2.6
Equation of the Tangent Line to a Function at a Point
The tangent line to the graph of a function f at a point x = c is defined to be the line
passing through (c, f (c)) with slope f (c), provided that the derivative f (c) exists. This
line has equation
y = f (c) + f (c)(x − c).
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
Formal Definition of the Derivative
173
Proof. We need only calculate the form of the line that passes through the point (c, f (c)) and has
slope f (c). Using the point-slope form y − y0 = m(x − x0 ), we see that this line has equation
y − f (c) = f (c)(x − c),
and solving for y, we obtain the desired equation y = f (c) + f (c)(x − c).
For example, we saw earlier that the derivative of f (x) = x 2 at x = 3 is equal to f (3) = 6.
This means that the tangent line to the graph of f (x) = x 2 at x = 3 has slope f (3) = 6
and passes through (3, f (3)) = (3, 9). This line has equation y = 9 + 6(x − 3), which, in
slope-intercept form, is the equation y = 6x − 9.
The tangent line to a function f at a point x = c is the unique line that “agrees with”
both the height of the function and the slope of the function at x = c. This means that near
x = c the graph of a function is very close to the graph of its tangent line. Therefore we can
use the tangent line as a rough approximation to the function f itself near x = c.
DEFINITION 2.7
Local Linearity
If f has a well-defined derivative f (c) at the point x = c, then, for values of x near c, the
function f (x) can be approximated by the tangent line to f at x = c with the linearization
of f around x = c given by
f (x) ≈ f (c) + f (c)(x − c).
Note that this definition does not assert how good an approximation one can make by using
the tangent line. What constitutes “near” and what constitutes “good” will be determined
by the context of the problem at hand.
For example, since the tangent line to f (x) = x 2 at x = 3 is the line y = 6x − 9, the line
y = 6x − 9 can be used as a rough approximation to the graph of the function f (x) = x 2 , at
least for values of x close to 3. This approximation will be better the closer we are to x = 3.
As the following figure shows, the two graphs have nearly the same height at the point
x = 3.5 and are still relatively close even at x = 4.
y = 6x − 9 is close to f (x) = x 2 near x = 3
y
24
21
18
15
12
9
6
3
2 1
1
2
3
4
x
The concept of local linearity will be particularly helpful for approximating roots with
Newton’s method in Example 8 and Exercises 81–86, as well as for approximating solutions of differential equations with Euler’s method in Section 7.5.
Leibniz Notation and Differentials
Derivatives are used in so many different fields of study that they are represented with a
wide variety of notations. In this book we will focus on two types: first, the “prime” notation
df
f (x) that we have already established, and second, the Leibniz notation . If y = f (x),
then we can also write y (x) or
dy
dx
dx
to represent the derivative.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
174
Chapter 2
November 21, 2012
Derivatives
Leibniz notation is named for one of the founders of calculus and is intentionally structured to remind us of the connection between derivatives and average rates of change. (The
“prime” notation we have been using so far is due to Lagrange.) Intuitively, the expression
dx represents an infinitesimally small change in x, just as x represents a small finite change
in x. In Leibniz notation the definition of the derivative as a limit of average rates of change
is strikingly clear:
dy
y
= lim
.
dx
x→0 x
It is sometimes convenient to think of the expressions dy and dx as differentials in a
functional relationship, with dy as a function of dx, in the following sense: We know that
y
represents the slope of a line with vertical change of y for a given horizontal change
x
x, as in the figure shown next at the left. In the same way we might try to think of
dy
dx
as
the slope of a line with a vertical change of dy for a given horizontal change dx, as in the
figure at the right.
y depends on x
and gives the height of the function
As differentials, dy depends on dx
and give the height of the tangent line
y
y
f(c dx)
f(c) dy
f(x)
y
f(c)
dy
f(c)
x
c
x
dx
x
c
x
x
With this interpretation of dy as a differential we can recast local linearity (Definition 2.7)
as saying that when x is sufficiently close to a point c, we can approximate f (x) by adding
dy to f (c). If x = c + dx as in the rightmost figure, this approximation can be expressed as
f (c + dx) ≈ f (c) + dy.
CAUTION
It is important to note that we will not be thinking of dx and dy as numbers, but rather
dy
thinking of
as a formal symbol that represents the derivative f (x). This formal symbol
dx
suggests meanings such as those given to the differentials in the preceding figure, but
dy
dx
is not an actual quotient of numbers; it is a limit of slopes of secant lines as defined in
Definition 2.2.
We can also write Leibniz notation in a slightly different way, called operator notation,
as follows:
dy
d
= ( y(x)).
dx
Here we are thinking of
d
dx
dx
as a sort of metafunction that operates on functions instead
of numbers: It takes functions as inputs and returns the derivatives of those functions as
dy
= 2x”
outputs. Using operator notation, we can express the statement “if y = x 2 , then
dx
d
dx
in the more compact form “ (x 2 ) = 2x.”
Although Leibniz notation is sometimes more convenient or informative than the usual
“prime” notation, it has one drawback: It is cumbersome to write down the point-derivative
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
175
Formal Definition of the Derivative
of a function in Leibniz notation. For example, as we showed earlier, the derivative of
d
f (x) = x 2 is f (x) = 2x. This derivative is easy to express in Leibniz notation as (x 2 ) = 2x.
dx
Now suppose we wish to consider the same derivative at the point x = 3. In “prime notation” it is easily expressed as f (3) = 2(3) = 6. In Leibniz notation it is more difficult to
work the x = 3 evaluation into the notation. We could write any of the following:
df d 2 d 2 (x )
= 6, or
(x ) = 6.
= 6,
dx 3
dx
dx
x=3
3
If f is a function, then its derivative f is a function as well. This means that we can also
consider its derivative, which we call f , the second derivative of f . In Leibniz notation we
d2f
. This notation is supposed to suggest the fact that we are
dx 2
d d
( f (x) . Similarly, we could find the third, fourth,
differentiating twice, that is, finding
dx dx
write the second derivative as
or fifth derivatives of a function f , and so on. For example, the third derivative of f (x) can
be written as f (x). For larger values of n we will replace the primes with a parenthetical
notation: for example, f (x) = f (6) (x). In general, the nth derivative of a function f is
denoted by:
f (n) (x) =
d nf
d d d
d
=
···
( f (x)) · · ·
dxn
dx dx dx
dx
.
n times
Examples and Explorations
EXAMPLE 1
Calculating the derivative at a point
Consider the function f (x) = x 3 .
(a) Use the h → 0 definition of the derivative to find f (2).
(b) Use the z → x definition of the derivative to find f (2).
(c) Find the equation of the tangent line to f (x) = x 3 at x = 2, and graph f (x) = x 3 and
this line on the same set of axes.
SOLUTION
(a) Using the h → 0 definition of the derivative, we have
f (2 + h) − f (2)
(2 + h)3 − 23
= lim
h
h
h→0
h→0
← derivative with x = 2
(8 + 12h + 6h2 + h3 ) − 8
h
h→0
← multiply out (2 + h)3
f (2) = lim
= lim
12h + 6h2 + h3
h(12 + 6h + h2 )
= lim
← algebra
h
h
h→0
h→0
= lim
= lim (12 + 6h + h2 ) = 12 + 6(0) + (0)2 = 12. ← cancellation, limit rules
h→0
(b) With the z → 2 definition of the derivative the algebra is different, but the final answer
will be the same. Along the way we will need to use the factoring formula a3 − b3 =
(a − b)(a2 + ab + b2 ), as follows:
f (z) − f (2)
z 3 − 23
= lim
z→2
z→2 z − 2
z−2
← derivative with x = 2
= lim
(z − 2)(z 2 + 2z + 4)
z→2
z−2
← factoring formula
= lim (z 2 + 2z + 4) = 22 + 2(2) + 4 = 12.
← cancellation, limit rule
f (x) = lim
z→2
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
176
Chapter 2
November 21, 2012
Derivatives
(c) We have just shown two ways that if f (x) = x 3 , then f (2) = 12. Thus the tangent line to
f (x) at x = 2 has slope 12 and passes through the point (2, f (2)) = (2, 8). The equation
of this line is
y − 8 = 12(x − 2) =⇒ y = 12x − 24 + 8 =⇒ y = 12x − 16.
When we graph the line along with the original function, we see that indeed the line
y = 12x − 16 is the tangent line to f (x) = x 3 at x = 2:
y = 12x − 16 is the tangent line to f (x) = x 2 at x = 2
y
16
12
8
4
3 2 1
4
1
2
3
x
8
EXAMPLE 2
Calculating the derivative of a function
Consider the function f (x) = x 3 .
(a) Use the h → 0 definition of the derivative to find f (x).
(b) Use the z → x definition of the derivative to find f (x).
(c) Graph the function f (x) = x 3 and the derivative f (x) you found in parts (a) and (b),
and argue that one graph is the slope function of the other.
SOLUTION
(a) The calculations in this example will be similar to those in the previous example, except
that we will not specify a specific value of x here. Using the h → 0 definition of the
derivative, we have
f (x + h) − f (x)
(x + h)3 − x 3
= lim
h
h
h→0
h→0
(x 3 + 3x 2 h + 3xh2 + h3 ) − x 3
= lim
h
h→0
3x 2 h + 3xh2 + h3
h(3x 2 + 3xh + h2 )
= lim
= lim
h
h
h→0
h→0
= lim (3x 2 + 3xh + h2 ) = 3x 2 + 3x(0) + (0)2 = 3x 2 .
f (x) = lim
h→0
← derivative
← multiply out (x + h)3
← algebra
← cancellation, limit rules
0
0
Notice that before we did any algebra, the limit was of the indeterminate form . For
derivative calculations with the h → 0 definition, the goal is often to expand and simplify until a common factor of h can be cancelled, as we did here.
(b) With the z → x definition of the derivative the algebra is different, but the final answer
is the same:
f (z) − f (x)
z3 − x3
f (x) = lim
= lim
← derivative
z→x
z→x z − x
z−x
= lim
z→x
(z − x)(z 2 + zx + x 2 )
z−x
← factoring formula
= lim(z 2 + zx + x 2 ) = x 2 + x 2 + x 2 = 3x 2 .
z→x
← cancellation, limit rules
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
177
Formal Definition of the Derivative
0
0
Again, notice that the limit was initially of the indeterminate form . For derivative
calculations with the z → x definition, the goal is often to factor and simplify until a
common factor such as z − x can be cancelled, as we just did.
(c) We have just shown two ways that if f (x) = x 3 , then f (x) = 3x 2 . When we graph
these two functions, we can see that at each value of x, the slope of the graph of
f (x) = x 3 is equal to the height of the graph of f (x) = 3x 2 . In particular notice that the
slopes of f (x) = x 3 are always positive and the graph of f (x) = 3x 2 is always positive.
...are heights of f (x) = 3x 2
Slopes for f (x) = x 3 ...
y
y
30
30
20
20
10
10
3 2 1
10
EXAMPLE 3
1
2
3
x
3 2 1
10
20
20
30
30
1
2
3
x
Finding the equation of a tangent line
2
x
Suppose f (x) = . Find the equation of the tangent line to f at x = 1.
SOLUTION
By definition,
the
tangent line to f at x = 1 has slope f (1) and passes through the point
(1, f (1)) = 1,
2
1
= (1, 2). Before we can find the equation of this line, we must calculate
(1).
the value of f
A lot of algebra will be needed before we can cancel a common factor of
h and solve the limit:
2
2
−
f (1 + h) − f (1)
f (1) = lim
= lim 1 + h 1
h
h
h→0
h→0
2 − 2(1 + h)
2 − 2 − 2h
1+h
= lim
= lim
h
h→0
h→0 h(1 + h)
= lim
h→0
−2h
−2
−2
= lim
=
= −2.
h(1 + h) h→0 1 + h
1+0
← derivative with x = 1
← algebra
← cancellation, limit rules
We now use the point-slope form of a line to find the equation of the line that has slope
f (1) = −2 and that passes through the point (1, f (1)) = (1, 2):
y − 2 = −2(x − 1) =⇒ y = −2(x − 1) + 2 =⇒ y = −2x + 4.
CHECKING
THE ANSWER
We can verify the reasonableness of the answer we just found by graphing the function
2
f (x) = and the line y = −2x + 4 on the same set of axes and checking that the line apx
pears to be tangent to the graph of f at the point (1, f (1)):
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
178
Chapter 2
November 21, 2012
Derivatives
f (x) =
2
and y = −2x + 4
x
y
4
2
4
2
2
4
x
2
4
EXAMPLE 4
Finding derivatives of functions that involve roots
d √
( x ).
dx
d √
find ( x ).
dx
(a) Use the z → x definition of the derivative to find
(b) Use the h → 0 definition of the derivative to
SOLUTION
(a) In the calculation that follows√we will use√ the factoring formula a2 − b2 = (a − b)
(a + b) in the case
√ where
√ √a = √z and b = x. In other words we will apply the formula z − x = ( z − x )( z + x ). Remember that our goal in this calculation of the
derivative is to factor and simplify until we can cancel something in the numerator
with the same thing in the denominator and take the limit:
√
√
z− x
d √
( x ) = lim
← derivative
dx
z→x
z−x
√
√
z− x
← factoring formula
= lim √
√ √
√
z→x ( z − x )( z + x )
= lim √
z→x
= √
1
√
z+ x
1
1
√ = √ .
x+ x
2 x
← cancellation
← take limit, algebra
(b) We encounter different algebra when we use the h →
√ 0 definition
√ of the derivative.
√
√
The conjugate of an expression of the form a − b is a + b, and vice versa.
Notice that the product of √
such an expression
and its conjugate does not involve any
√
√
√
square roots, since ( a − b )( a + b ) = a − b. In the calculation that follows we
will simplify our limit by multiplying numerator and denominator by a conjugate, to
clear square roots. Remember that our goal in this calculation is to cancel a common
factor of h so that we can take the limit:
√
√
x+h− x
d √
( x ) = lim
← derivative
dx
h
h→0
√
√
√
√
x+h− x
x+h+ x
← multiply by conjugate
= lim
√
√
h
h→0
x+h+ x
= lim
h→0
(x + h) − x
h
√
√ = lim √
√ ← algebra
h( x + h + x ) h→0 h( x + h + x )
= lim √
h→0
1
1
√ = √ .
√ = √
2 x
x+0+ x
x+h+ x
1
← cancel, take limit
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
EXAMPLE 5
Formal Definition of the Derivative
179
The derivative of a piecewise-defined function
Consider the piecewise-defined function f (x) =
3x + 2,
6 − x,
if x < 1
if x ≥ 1.
(a) Calculate f + (1) and f − (1). What can you say about f (1)?
(b) Write down a formula for f (x) as a piecewise-defined function.
SOLUTION
(a) We start by finding the right derivative of f at x = 1. In this case we examine h → 0+ ,
which means that h > 0, and thus 1 + h > 1. Therefore we will use the second part of
the piecewise-defined function f to evaluate f (1 + h) in this case:
f (1 + h) − f (1)
(6 − (1 + h)) − (6 − 1)
= lim
h
h
h→0+
6−1−h−6+1
−h
= lim
= lim
= lim (−1) = −1.
h
h→0+
h→0+ h
h→0+
f + (1) = lim
h→0+
In contrast, when we calculate the left derivative of f at x = 1, we will have h → 0− ,
and thus h < 0. This means that 1+h < 1, so we will use the first part of the piecewisedefined function f to evaluate f (1 + h). Of course we still have 1 ≥ 1, so we still use
the second part of f to evaluate f (1):
f (1 + h) − f (1)
(3(1 + h) + 2) − (6 − 1)
= lim
h
h
h→0−
3 + 3h + 2 − 6 + 1
3h
= lim
= lim
= lim 3 = 3.
h
h→0−
h→0− h
h→0−
f − (1) = lim
h→0−
Since the left and right derivatives of f at x = 1 are not equal to each other, the derivative f (1) of f at x = 1 is undefined. Note that f is left differentiable and right differentiable at x = 1, but not differentiable at x = 1.
(b) We have just calculated that f (1) does not exist. It now remains to determine f (x) for
values of x that are less than or greater than 1. For x < 1 the value of f (x) is equal to
3x + 2. Using the definition of the derivative, we see that for x < 1 we have
(3(x + h) + 2) − (3x + 2)
3x + 3h + 2 − 3x − 2
= lim
h
h
h→0
h→0
3h
= lim
= lim 3 = 3.
h→0 h
h→0
f (x) = lim
For x > 1 the value of f (x) is equal to 6 − x. A similar calculation shows that for x > 1
we have
(6 − (x + h)) − (6 − x)
6−x−h−6+x
= lim
h
h
h→0
−h
= lim
= lim −1 = −1.
h→0 h
h→0
f (x) = lim
h→0
Therefore the derivative of the piecewise-defined function f is
⎧
3, if x < 1
⎨
f (x) = undefined, if x = 1
⎩
−1, if x > 1,
where “DNE” stands for “does not exist,” representing the fact that the function f is
not defined at x = 1 (i.e., that there is no real number assigned to f (1)).
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
180
Chapter 2
CHECKING
THE ANSWER
November 21, 2012
Derivatives
The figure that follows at the left shows the graphs of the function f and its derivative f from the previous example. We can see from the graph of f that the right derivative of
the function at x = 1 is negative (think about the secant lines when h > 0) while the left
derivative at x = 1 is positive and steeper (think about the secant lines when h < 0). In fact,
since the pieces of this graph are linear, we can clearly see from the graph that the slopes
of the secant lines from the right and the left are −1 and 3, respectively.
y = f (x)
y = f (x)
y
1
y
5
5
4
4
3
3
2
2
1
1
1
1
2
3
x
1
1
1
2
3
x
Because the left and right derivatives at x = 1 do not agree, there is no well-defined tangent
line at x = 1; the (two-sided) derivative of f at x = 1 does not exist. Graphically, there is
no unique tangent line that passes through the point (1, f (1)). In some sense, the graph of f
has two directions at x = 1, one with slope −1 and one with slope 3. In the graph of f at
the right we see that the slopes of f are recorded correctly, with a slope of 3 for x < 1, a
slope of −1 for x > 1, and an undefined slope for x = 1.
EXAMPLE 6
A function that fails to be continuous also fails to be differentiable
Consider the piecewise-defined function f given by the equation
x 2,
12,
f (x) =
if x < 3
if x ≥ 3.
Argue graphically that f is not differentiable at x = 3 by examining secant lines on the
graph of f from both the left and the right of x = 3.
SOLUTION
The graph of f looks like this:
y
12
9
3
x
On the one hand, for h > 0, the secant lines over intervals of the form [3, 3 + h] all have
slope zero, since, on the interval [3, ∞), the graph of f is a horizontal line. Therefore the
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
181
Formal Definition of the Derivative
right derivative of f at x = 3 must be f + (3) = 0. On the other hand, for h < 0, the secant
lines over intervals of the form [3 + h, 3] behave as shown in the following three figures:
Secant line on [2, 3]
Secant line on [2.5, 3]
y
Secant line on [2.75, 3]
y
y
12
12
12
9
9
9
3
x
3
x
3
x
As h → 0− , we get a sequence of secant lines that are becoming more and more vertical.
The slopes of these secant lines are large and positive, and approach ∞ as h → 0− . Therefore the left derivative f − (3) does not exist. This means that f is not differentiable at x = 3,
since its left derivative does not exist.
EXAMPLE 7
Approximating roots with local linearity and Newton’s method
Use local linearity to approximate a root of f (x) = −x 3 − x + 1.
SOLUTION
First of all, notice that by the Intermediate Value Theorem and the fact that f (0) = 1 is
positive and f (1) = −1 is negative, the continuous function f (x) = −x 3 − x + 1 must have
at least one root between x = 0 and x = 1. However, since we cannot factor f (x), we are
unable to solve for the root directly. Our approximation strategy will be to guess a location
for a root, find the tangent line to the function at that point, and then look at the root of
that tangent line. If the tangent line is a good local approximation for the function, then
the root of the tangent line will be close to the root of the function if our initial guess is
close enough. We then repeat the process, using the root of the tangent line as our next
guess for the root of the function, to get better and better approximations of the root we
are looking for. This strategy is known as Newton’s method for approximating roots.
We’ll start with x1 = 0 as our first guess for a root of f (x). Clearly this is not actually a
root of the function, because f (0) = 1 is not equal to zero. Hopefully the tangent line at
x1 = 0 will help us make a better guess for a root. Before we compute the tangent line, we
must find the derivative of f (x) = −x 3 − x + 1. We will leave the bulk of the computational
details to the reader (see Exercise 42):
f (x + h) − f (x)
h
h→0
(−(x + h)3 − (x + h) + 1) − (−x 3 − x + 1)
= lim
h
h→0
2
= · · · = −3x − 1.
f (x) = lim
Therefore the derivative of f at x1 = 0 is f (0) = −3(0)2 − 1 = −1, and thus the tangent
line to f (x) at (0, f (0)) = (0, 1) has equation
y = f (0) + f (0)(x − 0) = 1 + (−1)(x − 0) = 1 − x.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
182
Chapter 2
November 21, 2012
Derivatives
Although it is not easy to calculate a root of our original cubic f (x) = −x 3 − x + 1, it is easy
to calculate that 1 is a root of the linear function y = 1 − x. Although x2 = 1 is still not a
root of f (x) (since f (1) = −1), the hope is that repeating this process will get us closer and
closer to a root. The following figure shows our initial guess of (x1 , f (x1 )) = (0, 1), the root
at x = 1 of the tangent line to f (x) at x1 = 0, and the new guess of (x2 , f (x2 )) = (1, −1).
f (x) = −x 3 − x + 1 with x1 = 0 and x2 = 1
y
2
1
0.5
1.0
1.5
x
1
2
Repeating this process, we can use a root of the tangent line to f (x) at x2 to obtain a
revised guess x3 for a root. The tangent line passes through (1, −1) and has slope f (1) =
−3(1) − 1 = −4, so its equation is
y = f (1) + f (1)(x − 1) = −1 + (−4)(x − 1) = −4x + 3.
The root of this line, and our new guess for a root of f (x), is x3 =
3
;
4
see the graph that
follows. Note that already our third approximation, x3 , appears quite close to the actual
root of f (x):
f (x) = −x 3 − x + 1 with x2 = 1 and x3 =
3
4
y
2
1
0.5
1.0
1.5
x
1
2
Repeating this process one more time, we obtain x4 ≈ 0.686047, which is in fact extremely
close to the actual root of x ≈ 0.682328. In Exercise 22 you will compare the accuracy of
this method with that of the Bisection Method of finding roots that we used in Exercise 80
of Section 1.4.
TEST YOUR
? UNDERSTANDING
What is the definition of the derivative of a function f at a point x = c? Give two an-
swers, one with a limit that involves h and one with a limit that involves z.
When calculating a derivative with the h → 0 definition of the derivative, why is the
0
0
limit always initially of the form ? What about when we use the z → x definition of
the derivative?
Is a function that is continuous at x = c necessarily differentiable at x = c? Is a function
that is differentiable at x = c necessarily continuous at x = c?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.2
What is the connection between the Leibniz notation
difference quotient
183
Formal Definition of the Derivative
y
?
x
dy
dx
for the derivative and the
How can we express the fifth derivative of a function f at the point x = 2 in Leibniz
notation? In “prime” notation?
EXERCISES 2.2
Thinking Back
Simplifying quotients: Simplify and rewrite the following expressions until you can cancel a common factor in the numerator and the denominator.
(x + h)−2 − x−2
(x + h)4 − x 4
h
h
z4 − x4
z−x
z−2 − x−2
z−x
Limit calculations: Find each of the following limits.
lim
h→0
lim
(1 − h) − 1
h
z→4
(3(−1 + h)2 + 1) − 4
h
lim
z2 − 4
z−2
h→0
1
− 0.5
2 + h
z→2
h
h→0
lim
lim
1
−1
lim z
(1 − 3z) + 11
z−4
z→1
z−1
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
f (x + h) − f (x)
.
h
f (x + h) − f (x)
.
(b) True or False: f (x) = lim
h
x→0
f (z) − f (x)
(c) True or False: f (x) = lim
.
z−x
z→0
(a) True or False: f (x) =
(d) True or False: If f (x) = x 3 , then f (x + h) = x 3 + h.
(e) True or False: If f (x) = x 3 , then f (x) =
lim
h→0
f (x 3 + h) − f (x)
.
h
(f) True or False: A function f is differentiable at x = c if
(c) and f +
(c) both exist.
and only if f −
(g) True or False: If f is continuous at x = c, then f is differentiable at x = c.
(h) True or False: If f is not continuous at x = c, then f is
not differentiable at x = c
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function that is continuous, but not
differentiable, at x = 2.
(b) The graph of a function that is left and right differentiable, but not differentiable, at x = 3.
(c) The graph of a function that is differentiable on the
interval [−1, 1] but not differentiable at the point
x = 1.
3. Use limits to give mathematical definitions for each of the
following derivatives, first with the h → 0 definition of
the derivative, and then with the z → x definition:
(a) the derivative of a function f at the point x = 5
(b) the derivative of a function f
(c) the right derivative of a function f at the point x = −2
4. Use limits to give mathematical definitions for:
(a) the slope of the line tangent to the graph of a function
f at the point x = 4
(b) the line tangent to the graph of a function f at the
point x = 4
(c) the instantaneous rate of change of a function f at the
point x = 1
(d) the acceleration at time t = 1.65 of an object that
moves with position function s(t)
5. Explain why the limits lim
h→0
f (x + h) − f (x)
f (z) − f (x)
and lim
z→x
h
z−x
are the same for any function f . (Hint: Consider the substitution z = x + h.)
f (x + h) − f (x)
f (z) − f (x)
and lim
z→x
h
z−x
0
are each initially in the form . Why would cancelling a
0
6. Explain why the limits lim
h→0
common factor of h or z − x be likely to resolve this indeterminate form?
7. The function f (x) = 4x 3 − 5x + 1 is both continuous
and differentiable at x = 2. Write these facts as limit
statements.
8. The function f (x) = 4 − x 2 is both continuous and differentiable at x = 1. Write these facts as limit statements.
9. If lim
x→c
f (x) − f (c)
exists, what can you say about the differx−c
entiability of f at x = c? What can you say about the continuity of f at x = c?
lim− f (x) = 1,
10. Suppose
f (0) = 1,
lim+ f (x) = 1,
x→0
x→0
f (x) − f (0)
f (x) − f (0)
= 3, and lim−
= −2.
x
x
x→0
x→0
lim+
(a) Is f continuous and/or differentiable at x = 0? What
about from the left or right?
(b) Sketch a possible graph of f .
11. Suppose
lim
h→0−
f (1) = 3,
lim f (x) = 3,
x→1−
lim f (x) = 3,
x→1+
f (1 + h) − f (1)
f (1 + h) − f (1)
= 2, and lim+
= 0.
h
h
h→0
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
184
Chapter 2
November 21, 2012
Derivatives
(a) Is f continuous and/or differentiable at x = 0? What
about from the left or right?
(b) Sketch a possible graph of f .
12. Consider the function f graphed here:
y
15. Express the sentence “the derivative of f (x) = 3x 2 − 1 at
x = −4 is −24” in each of the following notations:
(a) prime
(c) operator
(b) Leibniz
16. Express the sentence “the fourth derivative of f (x) =
3x 2 − 1 is equal to 0” in each of the following notations:
(a) prime
(c) operator
(b) Leibniz
Suppose f (x) = x 3 − 2x + 1, and let y = f (x). It can be shown
that f (x) = 3x 2 −2. Use this information to determine the expressions in Exercises 17 and 18. (Note: No differentiation will
be necessary, since the derivatives are given. These problems are
just a test of your ability to interpret derivative notation.)
3
2
1
dy df
,
,
dx dx
dy 18.
,
dx x=2
17.
1
2
3
x
(a) Sketch secant lines from (2, f (2)) to (2 + h, f (2 + h))
on the graph of f , for the following values of h:
h = 0.5, h = 0.25, h = 0.1, h = −0.5, h = −0.25, and
h = −0.1.
(b) Use the secant lines you sketched in part (a)
to
lim
graphically
h→0−
evaluate
f (2 + h) − f (2)
.
h
lim
h→0+
f (2 + h) − f (2)
h
and
d
dy
( f (x)), and
( f (x)).
dx
dx
df d
dy
( f (x)) , and
( f (2))
,
dx x=2 dx
dx
x=2
Express each of the following limit statements as delta–epsilon
statements:
(3 + h)2 − 9
19. f (3) = lim
h→0
h
2
x
−
9
20. f (3) = lim
x→3 x − 3
√
21. Suppose that you know that the derivative of f (x) = x
1
(c) Use your answer from part (b) to show that f is not
differentiable at x = 2.
13. Sketch secant lines on a graph of f (x) = |x|, and use them
to argue that the absolute value function is not differentiable at x = 0.
14. The two-sided symmetric difference approximation for
the slope of a tangent line (see Exercise 20 in Section 2.1)
can sometimes be misleading. Use a sequence of symmetric difference approximations to estimate the derivative of f (x) = |x| at x = 0. What does your sequence of approximations suggest about f (0)? Does this seem right?
is equal to f (x) = √ . Use this fact, local linearity, and
2 x
√
√
the fact that 4 = 2 to approximate the value of 4.1.
How
√ close is your approximation with the approximation
of 4.1 that you can find with a calculator? (Hint: Consider
the tangent line to f (x) at x = 4, and use it to approximate the
function nearby.)
22. Suppose you wish to find a root of f (x) = x 3 + x 2 + x + 1
in the interval [−3, 2]. Compare the accuracy of Newton’s
method as applied in Example 8 with the accuracy of the
Bisection Method used in Exercise 80 of Section 1.4 for
the same function. Which method gets closest to the root
in three iterations?
Skills
Use (a) the h → 0 definition of the derivative and then
(b) the z → c definition of the derivative to find f (c) for
each function f and value x = c in Exercises 23–38.
23. f (x) = x 2 , x = −3
25. f (x) =
1
, x = −1
x
24. f (x) = x 3 , x = 1
26. f (x) =
1
, x=2
x2
27. f (x) = 1 − x 3 , x = −1
28. f (x) = x 4 + 1, x = 2
29. f (x) = x 1/2 , x = 9
30. f (x) = x −1/2 , x = 9
31. f (x) =
x−1
, x=2
x+3
32. f (x) =
x − 3x
, x=0
x+1
2
33. f (x) = e x , x = 0
34. f (x) = 2e x
35. f (x) = sin x, x = 0
36. f (x) = cos x, x = 0
37. f (x) = tan x, x = 0
38. f (x) = sec x, x = 0
Use the definition of the derivative to find f for each function
f in Exercises 39–54.
39. f (x) = −2x 2
40. f (x) = 4 + x 2
41. f (x) = x 3 + 2
42. f (x) = −x 3 − x + 1
2
x+1
1
45. f (x) = 2
x
√
47. f (x) = 3 x
44. f (x) =
43. f (x) =
49. f (x) =
√
46. f (x) =
48. f (x) =
2x + 1
50. f (x) =
x−1
x+3
x3
53. f (x) =
x+1
52. f (x) =
51. f (x) =
54. f (x) =
2
1−x
1
x3
3
√
x
1
√
2x + 1
1
x2 − 1
x2 − 1
x2 − x − 2
Use the definition of the derivative to find the derivatives
described in Exercises 55–58.
55. Find
d
d2
d3
(2x 3 ),
(2x 3 ), and
(2x 3 ).
dx
dx 2
dx 3
56. Find
d
d2
d3
(2x 3 ) ,
(2x 3 ) , and
(2x 3 ) .
dx
dx 2
dx 3
3
1
−2
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
2.2
d 4f
d 4 f and 4 .
4
dx
dx 2
d 2f
d4
2
(3)
58. Given 2 = 2 − x , find f (4) and 4 ( f (x)).
dx
dx
Use the definition of the derivative to find the equations of the
lines described in Exercises 59–64.
57. Given f (3) (x) = 3x 2 + 1, find
61. The line tangent to the graph of y = 1 − x − x 2 at the
point (1, −1).
62. The line tangent to the graph of y = 4x + 3 at the point
(−2, −5).
73.
74.
63. The line that passes through the point (3, 2) and is parallel
75.
64. The line that is perpendicular to the tangent line to f (x) =
x 4 + 1 at x = 2 and also passes through the point (−1, 8).
For each function f graphed in Exercises 65–68, determine the
values of x at which f fails to be continuous and/or differentiable. At such points, determine any left or right continuity or
differentiability. Sketch secant lines supporting your answers.
y
65.
y
66.
3
2
2
1
1
⫺3 ⫺2 ⫺1
⫺3 ⫺2 ⫺1
1
2
3
x
⫺1
y
67.
1
2
3
⫺2
For each function f (x) and interval [a, b] in Exercises 81–86,
use the Intermediate Value Theorem to argue that the function
must have at least one real root on [a, b]. Then apply Newton’s
method to approximate that root.
y
3
2
3
79.
80.
68.
1
77.
⫺1
1
⫺3 ⫺2 ⫺1
76.
78.
x
1
70. f (x) = x 2/3 , x = 0
, x=0
x
f (x) = |x 2 − 4|, x = 2
72. f (x) = |x 2 − 4|, x = −2
x + 4, if x < 2
f (x) =
x=2
3x, if x ≥ 2,
x 2 − 3, if x < 3
x=3
f (x) =
x + 2, if x ≥ 3,
x 2 , if x ≤ 1
f (x) =
x=1
2x − 1, if x > 1,
x 2 , if x ≤ 1
x=1
f (x) =
2x + 4, if x > 1,
⎧
1
⎨
, if x = 0
x sin
x=0
f (x) =
x
⎩
0, if x = 0,
⎧
1
⎨ 2
x sin
, if x = 0
f (x) =
x=0
x
⎩
0, if x = 0,
1, if x rational
f (x) =
x=1
x + 1, if x irrational,
x 2 , if x rational
x=1
f (x) =
2x − 1, if x irrational,
69. f (x) =
71.
1
at x = −1.
x
185
In Exercises 69–80, determine whether or not f is continuous
and/or differentiable at the given value of x. If not, determine
any left or right continuity or differentiability. For the last four
functions, use graphs instead of the definition of the derivative.
59. The tangent line to f (x) = x 2 at x = −3.
60. The tangent line to f (x) = x 2 at x = 0.
to the tangent line to f (x) =
Formal Definition of the Derivative
x
2
⫺3 ⫺2 ⫺1
⫺2
⫺1
⫺3
⫺1
1
2
3
x
81.
82.
83.
84.
85.
86.
f (x) = x 2 − 5, [a, b] = [1, 3]
f (x) = x 2 − 2, [a, b] = [1, 2]
f (x) = x 3 − 3x + 1, [a, b] = [0, 1]
f (x) = x 3 − 3x + 1, [a, b] = [1, 2]
f (x) = x 3 + 1, [a, b] = [−2, 1]
f (x) = x 4 − 2, [a, b] = [1, 2]
Applications
87. In Example 4 of Section 2.1, we saw that a watermelon
dropped from a height of 100 feet will be s(t) = −16t 2 +
100 feet above the ground t seconds after it is dropped.
In that example, we approximated the velocity of such a
watermelon at time t = 1 by calculating a sequence of
average rates of change. Now we can calculate this velocity exactly, using the definition of the derivative. Do so,
and compare the exact answer to the approximation we
found earlier.
88. On a long road trip you are driving along a straight
portion of Route 188. Suppose that t hours after entering Nevada your distance from the Donut Hole is
s(t) = −10t 2 − 40t + 120 miles.
(a) How long will it take you to reach the Donut Hole
after entering Nevada?
(b) Find your velocity v(t) as you drive toward the Donut
Hole.
(c) Are you accelerating or decelerating as you approach
the Donut Hole? At what rate?
89. Suppose your position s(t) as you drive north along a
straight highway is as shown in the graph that follows
at the left, with t measured in hours and s measured in
miles.
(a) Sketch a graph of your velocity v(t), and use the graph
to describe your velocity over the course of the twohour drive.
CONFIRMING PAGES
23:44
TKmaster2010
WHF00153/FREE087-Taalman
186
Chapter 2
November 21, 2012
Derivatives
(b) Sketch a graph of your acceleration a(t), and use the
graph to describe your acceleration over the course of
the two-hour drive.
Position on highway
Carol’s velocity
y
y
100
200
75
100
50
10
25
20
30
x
100
1
2
x
200
90. While Carol is walking along a straight north-south forest path, her velocity, in feet per minute, after t minutes
is given by the preceding graph at the right. Suppose the
“positive” direction is north.
(a) Describe the sort of walk Carol must have taken to
have this velocity graph. Be sure your description explains the physical significance of the fact that her velocity is zero at t = 15 minutes and the fact that her
velocity is negative for the second half of her walk.
(b) Find Carol’s average acceleration over the 30-minute
walk. Was her acceleration constant over the duration
of her walk? Why or why not?
(c) What was Carol’s average velocity over her entire
walk? Why does your answer make sense?
91. To save up for a car, you take a job working 10 hours a
week at the school library. For the first six weeks the library pays you $8.00 an hour. After that you earn $11.50
an hour. You put all of the money you earn each week into
a savings account. On the day you start work your savings account already holds $200.00. Let S(t) be the function that describes the amount in your savings account t
weeks after your library job begins.
(a) Find the values of S(3), S(6), S(8), S (3), S (6), and
S (8), if possible, and describe their meanings in practical terms. If it is not possible to find one or more of
these values, explain why.
(b) Write an equation for the function S(t). (Hint: S(t) will
be a piecewise-defined function.) Be sure that your equation correctly produces the values you calculated in
part (a).
(c) Sketch a labeled graph of S(t). By looking at this
graph, determine whether S(t) is continuous and
whether S(t) is differentiable. Explain the practical
significance of your answers.
(d) Show algebraically that S(t) is a continuous function,
but not a differentiable function.
Proofs
92. Use the definition of the derivative to prove that every
quadratic function f (x) = ax 2 + bx + c has the property
that its graph has a horizontal tangent line at the point
x=−
b
.
2a
93. Use the definition of the derivative to prove that our concept of slope for linear functions matches the slope that
is defined by the derivative. In other words, show that if
f (x) = mx + b is any linear function, then f (x) = m.
94. Use Problem 93 to prove that a linear function is its own
tangent line at every point. In other words, show that if
f (x) = mx + b is any linear function, then the tangent line
to f at any point x = c is given by y = mx + b.
95. Use the mathematical definition of a tangent line and the
point-slope form of a line to show that if f is differentiable
at x = c, then the tangent line to f at x = c is given by the
equation y = f (c)(x − c) + f (c).
96. Prove that if a function f is differentiable at x = c, then f
is continuous at x = c.
(a) We are given that f is differentiable at x = c. Use the
alternative definition of the derivative to write down
what that statement means.
(b) We want to show that f is continuous at x = c.
Use the definition of continuity to show that
this statement is equivalent to the statement
lim ( f (x) − f (c)) = 0.
x→c
(c) Now use part (a) to show that lim ( f (x) − f (c)) = 0.
x→c
(Hint: Multiply ( f (x)−f (c)) by
rule for limits.)
x−c
and use the product
x−c
97. Use the definition of two-sided and one-sided derivatives, together with properties of limits, to prove that
(c) and f +
(c) exist and are
f (c) exists if and only if f −
equal.
98. Show that if a function y = f (x) is differentiable at x0 and
y = f (x0 + x) − f (x0 ),
then
y = f (x0 )x + x,
where is a function satisfying lim = 0.
x→0
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
187
Rules for Calculating Basic Derivatives
Thinking Forward
Derivatives of power functions: After differentiating a few power
functions, you may have noticed a pattern emerging. In the
following exercises we will investigate a possible formula for
differentiating power functions.
Derivatives of combinations of functions: We have already seen
that the limit of a sum is the sum of the limits and that the
limit of a product is the product of the limits. Do derivatives
also interact well with sums and products?
Use the z → x definition of the derivative to show that
d
(3x + x 2 ). Use your answers to make a condx
d
df
dg
jecture as to whether or not ( f (x) + g(x)) = + .
dx
dx
dx
d 8
(x ) = 8x 7 . (Hint: The following factoring formula
dx
will come in handy: If n is a positive integer, then
zn − xn = (z − x)(zn−1 + zn−2 x + zn−3 x 2 + · · · + z 2 xn−3 +
zxn−2 + xn−1 ).)
Use the definition of the derivative (or exercises done
previously in this section) to find (a)
(b)
Use the preceding two derivative formulas to make a
d
(x − 3),
dx
d
d
(2x + 1), and (c) ((x − 3)(2x + 1)). Use your
dx
dx
answers to make a conjecture as to whether or not
d
df
dg
.
( f (x)g(x)) =
d
conjecture about a formula for (xn ), where n is a posdx
itive integer.
2.3
d
d
(3x), (b) (x 2 ),
dx
dx
and (c)
Use the z → x definition of the derivative to show
that
Use the definition of the derivative (or exercises done
previously in this section) to find (a)
d 4
(x ) = 4x 3 .
dx
dx
dx
dx
RULES FOR CALCULATING BASIC DERIVATIVES
Formulas for differentiating constant, identity, linear, and power functions
Rules for differentiating constant multiples, sums, products, and quotients
Using differentiation rules to quickly calculate derivatives and antiderivatives
Derivatives of Linear Functions
You may have noticed by now that using the definition of the derivative to calculate derivatives can be rather tedious. Since derivatives will be used often throughout this course, we
need to develop a faster method of calculating them. Let’s start with linear functions. If f
is a linear function, then it has slope m everywhere, and therefore its derivative is constantly m. Since constant and identity functions are linear functions with slopes 0 and
1, their derivatives are constantly 0 and 1, respectively. These are our first differentiation
rules.
THEOREM 2.8
Derivatives of Constant, Identity, and Linear Functions
For any real numbers k, m, and b,
(a)
d
(k) = 0
dx
(b)
d
(x) = 1
dx
(c)
d
(mx + b) = m
dx
With this differentiation rule we can find derivatives of linear functions very quickly,
d
(3) = 0,
without having to consider the definition of the derivative. For example,
d
(3x
dx
+ 1) = 3, and
d
(2x
dx
dx
− 99) = 2.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
188
Chapter 2
November 21, 2012
Derivatives
Proof. We need only prove the third formula, since first two are special cases of that formula. Our
proof is just a general calculation with the definition of the derivative: If m and b are any constants
and f (x) = mx + b, then
f (x + h) − f (x)
h
← definition of the derivative
= lim
(m(x + h) + b) − (mx + b)
h
← use formula for f (x)
= lim
mx + mh + b − mx − b
h
← algebra
= lim
mh
= lim m = m.
h→0
h
← limit of a constant
f (x) = lim
h→0
h→0
h→0
h→0
The Power Rule
You may have already noticed a particular pattern for the derivatives of power functions. In
various examples and exercises in the previous section, we have seen that
d 3
d 4
d √
1
d
1
−3
= 4.
(x ) = 3x 2 ,
(x ) = 4x 3 ,
( x) = √ ,
and
3
dx
dx
dx
dx
2 x
x
x
The pattern becomes clear if we write these derivative formulas in exponent notation:
d 3
(x ) = 3x 2 ,
dx
d 4
(x ) = 4x 3 ,
dx
d 1/2
1
(x ) = x− 1/2 ,
dx
2
and
d −3
(x ) = −3x−4 .
dx
From these examples it appears that, to take the derivative of x k , we bring down the exponent k to the front of the expression and then decrease the exponent by one, to get kxk−1 .
This is in fact the case in general:
THEOREM 2.9
The Power Rule
For any nonzero rational number k,
d
(x k )
dx
= kx k−1 .
Although we require that k be a rational number in this formula, the power rule is actually
true for any nonzero real number k. If k = 0, then x 0 is the constant function 1, which we
already know how to differentiate.
Considering the algebra involved in applying the definition of the derivative, it is a relief
to have such a simple formula for finding derivatives of power functions! With this formula
d
d
we can quickly say, for example, that (x15 ) = 15x14 or that (x−1000 ) = −1000 x−1001 , or
dx
d
17 5/12
(x 17/12 ) =
x
. We can also find higher
dx
12
f (x) = x 4 , then f (x) = 4x 3 , f (x) = 12x 2 , f (x) =
even that
dx
derivatives very easily; for exam-
24x, f (4) (x) = 24, and f (5) (x) = 0.
ple, if
These functions are graphed next; each one is the associated slope function for the one
before.
f (x)
f (x)
f (x)
f (x)
f (4) (x)
f (5) (x)
y
y
y
y
y
y
x
x
x
x
CONFIRMING PAGES
x
x
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
Rules for Calculating Basic Derivatives
189
Proof. We will prove the power rule by cases. The case where k is a positive integer is proved
in what follows. The case where k is a negative integer is left for Exercise 86. The case where k is
p
a general rational number must be put off until after we know about implicit differentiation in
q
Section 2.4.
Given any positive integer k, we need to apply the definition of the derivative to find
d k
(x ). It
dx
turns out that the z → x definition of the derivative is easier to use in this case. Along the way we
will require the following factoring formula: If k is a positive integer, then
z k − x k = (z − x)(z k−1 + z k−2 x + z k−3 x 2 + · · · + z 2 x k−3 + zx k−2 + x k−1 ).
(You can verify this formula by simply multiplying out the right-hand side of the equation; everything except the first and last terms will cancel, and you will obtain the left side of the equation.)
Now using the definition of the derivative, we have
d k
zk − xk
(x ) = lim
z→x z − x
dx
(z − x)(z k−1 + z k−2 x + z k−3 x 2 + . . . + zx k−2 + x k−1 )
= lim
z→x
z−x
← definition of the derivative
← factoring formula
= lim(z k−1 + z k−2 x + z k−3 x 2 + . . . + zx k−2 + x k−1 )
← cancellation
z→x
=x
k−1
=x
k−1
= kx
+x
k−2
x+x
+x
k−1
+x
k−1
x + . . . + xx
k−3 2
k−1
+ ... + x
k−1
k−2
+x
+x
k−1
.
← evaluate limit
← there are k of these
k−1
.
The Constant Multiple and Sum Rules
In Chapter 1 we saw that limits behave well with respect to arithmetic combinations of
functions. For example, limits commute with sums: The limit of a sum is the sum of the
limits, provided that the limits involved exist. Is the same thing true for derivatives? For
constant multiples and sums, the answer is yes:
THEOREM 2.10
Derivatives of Constant Multiples and Sums of Functions
If f and g are functions and k is a constant, then for all x where the functions involved
are differentiable, we have the following differentiation formulas:
Constant Multiple Rule: (kf ) (x) = kf (x)
Sum Rule: ( f + g) (x) = f (x) + g (x)
Difference Rule: ( f − g) (x) = f (x) − g (x)
The notation (kf ) (x) indicates that we are differentiating the function kf with respect to x.
d
In Leibniz notation we would write this differentiation as (kf (x)). The difference rule is
dx
of course just a combination of the sum and constant multiple rules, since f (x) − g(x) =
f (x) + (−g(x)).
These rules mean that we can factor out constants and split up sums when calculating derivatives. For example, derivatives of sums or constant multiples of the functions
f (x) = x 2 and g(x) = x 4 can be expressed as sums or constant multiples of their derivatives
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
190
Chapter 2
November 21, 2012
Derivatives
f (x) = 2x and g (x) = 4x 3 , as illustrated here:
d
d
d
(5x 2 + 2x 4 ) = (5x 2 ) + (2x 4 )
dx
dx
dx
=5
← sum rule
d 2
d
(x ) + 2 (x 4 )
dx
dx
← constant multiple rule
= 5(2x) + 2(4x 3 )
← power rule
= 10x + 8x 3 .
Proof. We will use the definition of the derivative to prove (a) the constant multiple and (b) sum
rules in what follows. The difference rule can also be proved with the definition of the derivative,
or by applying the sum and constant multiple rules; see Exercise 87.
(a) The proof of the constant multiple rule uses the definition of the derivative and the constant
multiple rule for limits. Given a function f (x) and a constant k, we wish to show that the derivative of the function kf (x) is the same as k times the derivative of f (x). We’ll start from the left
and work to the right:
(kf ) (x) = lim
h→0
= lim
h→0
kf (x + h) − kf (x)
h
← definition of the derivative for kf (x)
k( f (x + h) − f (x))
h
← factor out k from numerator
f (x + h) − f (x)
= lim k
h
h→0
=k
lim
h→0
f (x + h) − f (x)
h
← factor out k from quotient
← constant multiple rule for limits
= kf (x).
← definition of the derivative for f (x)
(b) Similarly, the proof of the sum rule uses the definition of the derivative and the sum rule for
limits. We wish to show that the derivative of the function f + g is the sum of the derivative of f
and the derivative of g. We will work from the left to the right; our goal is to use the definition
of the derivative to write the left-hand statement as a limit and then use algebra and the sum
rule for limits to split this limit into two limits, one of which will be the derivative of f , and one
the derivative of g:
( f (x + h) + g(x + h)) − ( f (x) + g(x))
h
← definition of derivative
= lim
f (x + h) + g(x + h) − f (x) − g(x)
h
← simplify
= lim
( f (x + h) − f (x)) + ( g(x + h) − g(x))
h
( f + g) (x) = lim
h→0
h→0
h→0
= lim
h→0
=
lim
h→0
f (x + h) − f (x)
g(x + h) − g(x)
+
h
h
f (x + h) − f (x)
h
= f (x) + g (x).
+ lim
h→0
← reordering terms
g(x + h) − g(x)
h
← algebra
← sum rule for limits
← definition of derivative
The Product and Quotient Rules
We now know that derivatives interact nicely with constant multiples and sums; for
example, the derivative of a sum is the sum of the derivatives. Do derivatives also commute with products and quotients? Sadly, the answer is no. For example, consider the
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
Rules for Calculating Basic Derivatives
191
functions f (x) = x 2 + 1 and g(x) = x 3 , whose derivatives are f (x) = 2x and g (x) = 3x 2 . Now
consider the product f (x)g(x) = (x 2 + 1)(x 3 ) = x 5 + x 3 ; by the power and sum rules, the
derivative of that product is 5x 4 + 3x 2 . However, this is clearly not equal to the product
f (x)g (x) = (2x)(3x 2 ) = 6x 3 .
In general, the derivative of a product is not necessarily the product of its component
derivatives. Similarly, the derivative of a quotient is not in general the quotient of its comf
ponent derivatives. However, we can write the derivative of a product fg or quotient in
terms of f , g, f , and g , with the following, somewhat surprising, formulas:
THEOREM 2.11
g
Derivatives of Products and Quotients of Functions
If f and g are functions, then for all x such that both f and g are differentiable, we have
the following differentiation formulas:
Product Rule: ( fg) (x) = f (x) g(x) + f (x) g (x)
Quotient Rule:
f
g
(x) =
f (x) g(x) − f (x)g (x)
( g(x))2
Some people remember the product rule by remembering the pattern of differentiating
one piece and not the other in both possible ways and then taking the sum. Some people
remember the quotient rule with a phrase like “lo d-hi minus hi d-lo over lo lo,” although
it might be easier just to remember the quotient rule than to try and remember that! In
any case, you will have to memorize these differentiation formulas, because they will be
needed often throughout this book.
Armed with the product and quotient rules, we can now differentiate a lot more functions. For example, with our earlier example of f (x) = x 2 + 1 and g(x) = x 3 , we can differentiate f (x)g(x) without multiplying it out first:
d
( f (x)g(x)) = f (x) g(x) + f (x) g (x)
dx
← product rule
= (2x)(x 3 ) + (x 2 + 1)(3x 2 )
← f (x) = 2x, g(x) = 3x 2
= 2x 4 + 3x 4 + 3x 2
← algebra
= 5x 4 + 3x 2 .
This is exactly what we found earlier by multiplying out f (x)g(x) first and then taking the
derivative.
Proof. The proof of the product rule is yet another general calculation using the definition of the
derivative, with a small twist that we will explain shortly. A similar method of proof works for the
quotient rule and is left to Exercise 89. We will have an easier way to prove the quotient rule once
we learn the chain rule in Section 2.4.
Suppose f and g are differentiable functions. We wish to show that ( fg) = f g + fg . We will
apply the definition of the derivative to the left side of that equation and then use a long string
of algebra and limit rules to rewrite it in terms of f , g, f , and g . This will be made possible in the
second step of the following calculation by adding and subtracting the same expression, f (x)g(x+h),
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
192
Chapter 2
November 21, 2012
Derivatives
from the numerator.
f (x + h)g(x + h) − f (x)g(x)
h
f (x + h)g(x + h) − f (x)g(x + h) + f (x)g(x + h) − f (x)g(x)
= lim
h
h→0
( f (x + h) − f (x))g(x + h) + f (x)( g(x + h) − g(x))
= lim
h
h→0
f (x + h) − f (x)
g(x + h) − g(x)
= lim
· g(x + h) + f (x) ·
h
h
h→0
( fg) (x) = lim
← derivative
h→0
h→0
lim
f (x + h) − f (x)
h
h→0
h
g(x + h) − g(x)
h
← algebra
← factoring
← algebra
lim g(x + h) + lim f (x) lim
h→0
h→0
f (x + h) − f (x)
g(x + h) − g(x)
g(x) + f (x) lim
= lim
← limit rules
= f (x) g(x) + f (x) g (x).
← derivative
=
h→0
h
h→0
← continuity
The step labeled “continuity” follows from the fact that g(x) is differentiable and therefore continuous, because that is what allows us to say that lim g(x + h) = g(x). Notice that throughout the
h→0
entire calculation, our goal was to extract the expressions that represent the derivatives of f and g.
The algebra steps were not meant to “simplify”; they were meant to get us closer to the final form
of f g + fg .
Examples and Explorations
EXAMPLE 1
Determining whether the differentiation rules apply
Find the derivatives of each of the functions that follow, using the differentiation rules from
this section, if possible. If it is not possible, explain why not.
√
(a) f (x) = 2(5 x )
(b) g(x) = 34
(c) h(x) = (2x + 3)2
(d) k(x) = 3x 2 + 1
SOLUTION
(a) We cannot differentiate f (x) = 2(5 x ) with any of the rules from this section. In particular, the power rule applies only to power functions of the form x k , where the variable x
is in the base and a constant k is in the exponent. The function 5 x is not a power function, but an exponential function, which we will see how to differentiate in Section 2.5.
(b) Since g(x) = 34 = 81 is a constant function,
d 4
(3 )
dx
= 0.
(c) h(x) = (2x + 3)2 is a composition of functions, and we do not yet have a rule for
differentiating compositions. Therefore we must use algebra to expand the function
first. We could write the function as (2x + 3)(2x + 3) and use the product rule, or we
could just expand (2x + 3)2 entirely and then need only the sum and constant multiple
rules. We present the latter approach here:
d
d
((2x + 3)2 ) = (4x 2 + 12x + 9)
dx
dx
=
← algebra
d
d
d
(4x 2 ) + (12x) + (9)
dx
dx
dx
=4
d 2
d
d
(x ) + 12 (x) + (9)
dx
dx
dx
= 4(2x 1 ) + 12(1) + 0 = 8x + 12.
← sum rule
← constant multiple rule
← power rule
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
Rules for Calculating Basic Derivatives
193
√
(d) The function k(x) = 3x 2 + 1 is a composition that cannot be simplified, and we
do not yet have a rule for
√ differentiating compositions. (Note in particular that this
function is not equal to 3x + 1, because square roots do not distribute over sums;
there is no useful simplification that can be done to this function.) We will see how to
differentiate such functions in Section 2.4.
EXAMPLE 2
Finding derivatives with and without simplifying first
Differentiate each of the following functions with and without the quotient rule:
√
4
x7 + x
x2
(a) f (x) = 2
(c) h(x) =
(b) g(x) = −3 √
3x
x3
x
x
SOLUTION
(a) Applying the quotient rule gives us
4
d
dx 3x 2
=
d
d
(4) · (3x 2 ) − (4) · (3x 2 )
dx
dx
(3x 2 )2
← quotient rule
(0)(3x 2 ) − (4)(6x)
← constant and power rules
9x 4
8
−24x
= − x−3 .
← algebra
=
3
9x 4
Alternatively, we could simplify f first and then apply the constant multiple and power
rules. We of course will get the same final answer:
4
d
d 4 −2
← algebra
x
=
dx 3x 2
dx 3
=
=
4 d −2
(x )
3 dx
4
3
← constant multiple rule
8
3
= (−2x−3 ) = − x−3 .
← power rule, algebra
(b) If we apply differentiation rules immediately without simplifying first, we need to use
both the quotient and product rules:
√
√
d 2
d
(x ) · (x−3 x ) − (x 2 ) · (x−3 x )
x2
d
dx
dx
← quotient rule
=
√
√
dx x−3 x
(x−3 x )2
√
√
1
(2x)(x−3 x ) − (x 2 ) −3x−4 x + x−3 x− 1/2
2
. ← more rules
=
√
(x −3 x )2
However, if we do some preliminary algebra, then the differentiation steps and final
simplification will both be much, much simpler:
x2
d
d
← algebra
= (x 2 x 3 x− 1/2 )
√
dx x−3 x
dx
=
d 9/2
(x )
dx
← algebra
=
9 7/2
x .
2
← power rule
In Exercise 17 you will show that the preceding two answers are equal. Students often
want to differentiate as soon as possible, since the differentiation rules can be easier
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
194
Chapter 2
November 21, 2012
Derivatives
to implement than algebraic simplification. However, in many cases, in the long run,
simplifying before differentiating ends up saving a lot of even messier algebra later on.
Look at the end result of the two calculations we just did, and ask yourself which one
you would rather work with!
(c) Applying the quotient rule first, we have
√
√
d 7
d
7 √ (x + x ) · (x 3 ) − (x 7 + x ) · (x 3 )
d x + x
dx
dx
← quotient rule
=
dx
x3
(x 3 )2
√
1
7x 6 + x− 1/2 (x 3 ) − (x 7 + x )(3x 2 )
2
. ← sum, power rules
=
x6
Alternatively, we could do some algebra before differentiating and then apply the sum
and power rules. This gives us a simpler but equivalent answer:
7 √ 7
x 1/2
d x + x
d x
+
=
← algebra
dx
dx x 3
x3
x3
=
d 4
(x + x− 5/2 )
dx
← algebra
5
2
= 4x 3 − x− 7/2 .
← sum, power rules
Doing algebra before differentiating almost always means that there is much less algebra needed to simplify your answer after differentiating. In Exercise 18 you will show
that the two answers we just found are in fact equal.
CHECKING
THE ANSWER
You can always check the reasonableness of the answer to a differentiation problem by
graphing both f and f and verifying that f appears to be the associated slope function for
f . Do this for the following graphs, for part (c) of the example:
h(x) =
√
x7 + x
3
x
5
2
h (x) = 4x 3 − x− 7/2
100
100
0
0
3
100
0
EXAMPLE 3
Differentiating a piecewise-defined function
⎧
⎨ x 2 , if x ≤ 1
Find the derivative of the function f (x) =
⎩ 1 , if x > 1.
x
SOLUTION
For x < 1 we use the first case of the function to find the derivative:
f (x) =
d 2
(x ) = 2x.
dx
CONFIRMING PAGES
3
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
195
Rules for Calculating Basic Derivatives
Similarly, for x > 1 we use the second case:
d 1
d
−1
f (x) =
= (x−1 ) = −x−2 = 2 .
dx x
dx
x
It now remains only to determine what the derivative is at the breakpoint x = 1, if it exists.
The function f will be differentiable at x = 1 if it is continuous at x = 1 and the derivatives
1
of each of its pieces are equal at x = 1. Since (1)2 = , f is continuous at x = 1. However,
1
−1
,
the
derivatives
of
the
left
and
right
pieces
are not equal at x = 1. Therefore
(1)2
differentiable at x = 1, that is, f (1) does not exist. Thus the derivative of f is
since 2(1) =
f is not
f (x) =
⎧
⎪
⎪
⎨
2x,
if x < 1
DNE,
⎪
⎪
⎩ −1 ,
2
if x = 1
x
CHECKING
THE ANSWER
if x > 1.
Consider the graphs of f and f that follow. Note that f is indeed continuous, but not
differentiable, at x = 1. In the graph of f we have a hole at x = 1. We can see in both graphs
the fact that f − (1) = 2 but f + (1) = −1. We can also see that, except at x = 1, the slopes
of f (x) do seem to be the same as the heights of f (x).
f is undefined at x = 1
f is not differentiable at x = 1
y
y
2
2
1
2
1
1
2
3
x
1
1
EXAMPLE 4
1
2
3
x
2
Finding antiderivatives
An antiderivative of a function g(x) is a function whose derivative is g(x). In this problem
you may assume that any two functions with the same derivative must differ by a constant.
(We will prove this fact in Section 3.2.)
(a) Find one antiderivative of g(x) = 8x 3 .
(b) Describe the set of antiderivatives g(x) = 8x 3 .
(c) Find the one antiderivative of g(x) = 8x 3 that passes through (1, 4).
SOLUTION
(a) We can use a targeted guess-and-check method to find an antiderivative of g(x) = 8x 3 .
Since differentiating a power function decreases its power by one, we might start with
the function f (x) = x 4 , whose derivative is f (x) = 4x 3 ; this is almost what we want,
but off by a factor of two. A good second guess would be f (x) = 2x 4 , and indeed its
derivative is f (x) = 8x 3 , as desired.
(b) We now have one function whose derivative is f (x) = 8x 3 . Since any two functions
with that derivative must differ by a constant, the functions whose derivative is
f (x) = 8x 3 are all the functions of the form f (x) = 2x 4 + C for some real number
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
196
Chapter 2
November 21, 2012
Derivatives
C, as shown in the next figure at the left. The graph corresponding to our previous
choice of f (x) = 2x 4 , with C = 0, is shown in black.
One antiderivative has f (1) = 4
One antiderivative has C = 0
y
1
y
5
5
4
4
3
3
2
2
1
1
1
1
x
1
x
1
1
(c) Only one of the antiderivatives of f (x) = 8x 3 passes through the point (1, 4), as shown
in red at the right. To find this antiderivative we set f (1) = 4 and solve for C:
f (1) = 4 =⇒ 2(1)4 + C = 4 =⇒ C = 2.
Therefore the only antiderivative of f (x) = 8x 3 that passes through (1, 4) is the function f (x) = 2x 4 + 2.
Why are the rules for differentiating constants and the identity function special cases
TEST YOUR
? UNDERSTANDING
of the rule for differentiating a linear function?
What is the difference between the rule for differentiating a constant function and the
constant multiple rule?
Can you find examples of functions f and g such that the derivative of their quotient is
not the same as the quotient of their derivatives?
Why do we add and subtract f (x)g(x + h) in the calculation in the proof of the product
rule? What does that enable us to do in the calculation?
What sorts of functions can we differentiate with the rules in this section? Are there
any functions that we can’t differentiate with these rules?
EXERCISES 2.3
Thinking Back
Factoring formulas: Pull out a linear factor from each of the
following expressions.
z 2 − 100
z 3 − 27
Associated slope functions: For each of the following two function graphs, sketch a careful, labeled graph of its associated
slope function.
z 6 − 64
y
f (x) = x
4
f (x) = x 2 (x + 1)
f (x) = x
3
3
−2
x2
f (x) =
x+1
y
4
Definition-of-derivative calculations: Use the definition of the
derivative to find f for each function.
2
2
1
1
10
10
x
3 2 1
CONFIRMING PAGES
1
2
3
x
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.3
Rules for Calculating Basic Derivatives
197
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False:
(b) True or False:
(c) True or False:
(d) True or False:
(e) True or False:
d
(5) = 0.
dx
d
(ks + r) = k.
dr
d
(ks + r) = k.
ds
d
(3x + 1) k = k(3x + 1) k−1 .
dx
d 1
1
= 2.
dx x 3
3x
(f) True or False: If f and g are differentiable functions,
then ( f (x)g(x)) = g (x)f (x) + f (x)g(x).
(g) True or
If g and h are differentiable functions,
False:
then
g(x)
h(x)g (x) − g(x)h (x)
=
.
h(x)
(h(x))2
(h) True or False: Proving the sum rule for differentiation
involves the definition of the derivative, a lot of algebraic manipulation, and the sum rule for limits.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Functions f and g, which illustrate that, in general,
derivatives and products do not commute.
(b) Functions f and g, which illustrate that, in general,
derivatives and quotients do not commute.
(c) Three functions whose derivatives we cannot calculate by using the differentiation rules we have developed so far.
3. Express the constant multiple, sum, and difference rules
in Leibniz/operator notation.
4. Express the product and quotient rules in Leibniz/
operator notation.
5. Why does it make graphical sense that the derivative of a
constant is zero? That the derivative of the identity function is constantly equal to 1? That the derivative of a linear
function f (x) = mx + b is equal to m?
6. Why does the proof of the power rule (Theorem 2.9) in
this text work only when k is an integer? Also, at which
point in the proof do we use the fact that k is positive?
7. Explain why the power rule does not say that the derivative of 3 x is x3x−1 . Specifically, why doesn’t 3 x fit the pattern required for the power rule to apply?
8. Explain why the power rule cannot be used to differentiate the function (2 − x) 1/3 .
9. Use the product rule to find and state the rule for differentiating a product of three functions f , g, and h. In
.
other words, fill in the blank:( f (x)g(x)h(x)) =
Then use your rule to differentiate the function y =
(2x − 1)(x 2 + x + 1)(1 − 3x 4 ). Check your answer by differentiating the function y(x) another way.
10. Two operations commute if they can be done in either order. Does multiplying a function by a constant commute
with differentiation? Does adding two functions commute with differentiation? What about products and quotients and the operation of differentiation?
Given that f , g, and h are functions with values f (2) = 1,
g(2) = −4, and h(2) = 3 and point-derivatives f (2) = 3,
g (2) = 0, and h (2) = −1, calculate
11. (3f + 4h) (2)
13. ( f h) (2)
12. (2f + 3g − h) (2)
f
(2)
14.
g
15. In the text of this section we displayed graphs of f (x) = x 4
and its first five derivatives. Use the slope-height behavior of the graphs to verify that each is the associated slope
function of the one before.
16. Sketch graphs of f (x) =
1
and its first five derivatives.
x
Then use the slope-height behavior of the graphs to verify that each is the associated slope function of the one
before.
17. In Example 2(b) we calculated a derivative two different
ways. Use algebra to simplify the first answer and show
that it is equal to the second.
18. In Example 2(c) we calculated a derivative two different
ways. Use algebra to simplify the first answer and show
that it is equal to the second.
19. Suppose f is a piecewise-defined function, equal to g(x)
if x < 2 and h(x) if x ≥ 2, where g and h are continuous and differentiable everywhere. If g(2) = h(2), is the
function f necessarily differentiable at x = 2? Why or why
not?
20. Suppose f is a piecewise-defined function, equal to g(x) if
x < 2 and h(x) if x ≥ 2, where g and h are continuous and
differentiable everywhere. If g (2) = h (2), is the function
f necessarily differentiable at x = 2? Why or why not?
21. If possible, find constants a and b so that the function f
that follows is continuous and differentiable everywhere.
If it is not possible, explain why not.
3x + a, if x < 1
f (x) =
x b/2 , if x ≥ 1.
22. If possible, find constants a and b so that the function f
that follows is continuous and differentiable everywhere.
If it is not possible, explain why not.
ax − b, if x < 2
f (x) =
bx 2 + 1, if x ≥ 2.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
198
Chapter 2
November 21, 2012
Derivatives
Skills
Suppose g(x), h(x), and j(x) are differentiable functions with
the values of the function and its derivative given in the following table:
x
−1
0
1
2
3
g(x)
3
2
0
−2
−3
h(x)
0
3
−1
−2
0
j(x)
1
0
−2
−3
1
g (x)
−1
−2
−2
−1
0
h (x)
−2
3
−2
0
1
j (x)
−2
−2
−1
2
2
Use the table to calculate the values of the derivatives listed in
Exercises 23–28.
23.
24.
25.
26.
If f (x) = 3j(x) − 2h(x), find f (3).
If f (x) = g(x)h(x), find f (0).
If f (x) = g(x)(h(x) + j(x)), find f (2).
If f (x) = h(x)g(x)j(x), find f (1).
27. If f (x) =
3h(x)
, find f (0).
g(x) + j(x)
x2 − 1
x+1
49. f (x) = π 2
47. f (x) =
x2
51. f (x) = √
5
x2
√
5
x 7 − 2x 4
53. f (x) =
x3
x 7 − 3x 5 + 4
55. f (x) =
1 − 3x 4
1 3
1
57. f (x) =
+
√
x3
x
2x − 3
59. f (x) =
5x + 4
1
61. f (x) =
(x − 2)(x − 3)
x2
63. f (x) = 3
x + 5x 2 − 3x
x 4 − 7x 3
2x
√
√
50. f (x) = ( x − 3 x )2
48. f (x) =
52. f (x) =
x−1
(x + 1)(x + 2)
√
54. f (x) = (3x x )−2
x3 + x − 1
x2 − 7
x 2 − 3x
58. f (x) = 2
x − 2x + 1
√
60. f (x) = x 3 x (x 2/3 )
56. f (x) =
1
(x + 1)3
(x − 2)2
64. f (x) = 2
(x + 1)(x − 3)
62. f (x) =
Find the derivatives of each of the absolute value and
piecewise-defined functions in Exercises 65–72.
g(x)h(x) + j(x)
, find f (−1).
28. If f (x) =
h(x)
Differentiate each of the functions in Exercises 29–34 in two
different ways: first with the product and/or quotient rules and
then without these rules. Then use algebra to show that your
answers are the same.
3x + 1
29. f (x) = x 2 (x + 1)
30. f (x) =
x4
√
7/2
3
31. f (x) = x (2 − 5x )
32. f (x) = x (x−1 + 1)
√
x
x2 − x3
34. f (x) = −2 3
33. f (x) = √
x x
x
Use the differentiation rules developed in this section to find
the derivatives of the functions in Exercises 35–64. Note that it
may be necessary to do some preliminary algebra before differentiating.
65. f (x) = |x|
66. f (x) = |3x + 1|
67. f (x) = |1 − 2x|
68. f (x) = |x 2 − 1|
69. f (x) =
x 3 , if x < 1
x, if x ≥ 1
70. f (x) =
1, if x ≤ −1
x 2/3 , if x > −1
71. f (x) =
−x 2 , if x ≤ 0
x 2 , if x > 0
72. f (x) =
3x + 1, if x ≤ 1
x 3 , if x > 1
35. f (x) = 4 − 3x 7
36. f (x) = 2x − 3 + 4x 2
In Exercises 73–78, find a function that has the given derivative and value. In each case you can find the answer with an
educated guess-and-check process. It may be helpful to do
some preliminary algebra.
37. f (x) = 2(1 + 3x 2 )
38. f (x) = 5x 3 − 2x 2 + 7
73. f (x) = 3x 5 − 2x 2 + 4, f (0) = 1
39. f (x) = 2(3x+1)−4x 5
40. f (x) = x 2 + x(2 − 3x 2 )
41. f (x) = (x + 2)(x − 1)
42. f (x) = (x 2 − 3)2
43. f (x) = (3x + 2)3
44. f (x) = (3 − x)2 + 5
45. f (x) =
1 − 6x 3
3
46. f (x) =
x
x+1
74. f (x) = 7x 2 + 8x11 − 18, f (0) = −2
75. f (x) = 1 − 4x 6 , f (1) = 3
76. f (x) = x(4 − 2x), f (0) = 0
77. f (x) = (x 4 − 8)(1 − 3x 5 ), f (0) = 2
78. f (x) = (3x + 1)3 , f (2) = 1
Applications
79. A spaceship is moving along a straight path from Venus
into the heart of the Sun. The velocity of the spaceship t
hours after leaving Venus is v(t) = 0.012t 2 + 400 thousands of miles per hour. (To simplify matters we will pretend that Venus is not moving with respect to the sun;
you may assume that everything is fixed in place in this
exercise.)
(a) Say what you can about the initial values s0 , v0 , and
a0 , and then use derivatives and antiderivatives to
find equations for the position and acceleration of the
spaceship.
(b) Is the spaceship always moving towards the sun?
How can you tell?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
2.3
(c) Is the spaceship travelling at a constant acceleration?
Is it speeding up or slowing down, or neither? How
can you tell?
(d) The distance between Venus and the sun is about
67 million miles. How long will it take the spaceship
to reach the sun? How fast will the spaceship be going when it gets there?
t, hours
Venus
v(t)
t)
Sun
t)
a(t)
s0
x
s(t)
80. A bowling ball is thrown down from a 20th-story window. After 3 seconds, the bowling ball is 26 feet from
the ground and falling at a rate of −106 feet per second
(downwards). You may assume that gravity causes a constant downward acceleration of −32 feet per second.
Rules for Calculating Basic Derivatives
the position and velocity of the falling object after t
seconds are, respectively, s(t) = −16t 2 + v0 t + s0 and
v(t) = −32t + v0 .
82. In a science fiction novel, gravity on the planet XV-37 acts
very differently than it does in our universe. An object
dropped from a 1000-foot building on planet XV-37 will
have a downward gravitational acceleration of a(t) = −6t
feet per second per second after falling for t seconds. Use
antiderivatives to find equations for the position and velocity of such a falling object. What might be the consequences of living on this strange planet?
83. In another science fiction novel, gravity on planet Xillian again acts very strangely. The height s(t) of a falling
object on Xillian is always a cubic polynomial function.
Suppose a kiwi fruit is dropped (with initial velocity of
zero) from the top of a Xillian radio tower. After 5 seconds, the kiwi fruit is 100 feet from the ground and falling
at a rate of −200 feet per second. The acceleration of
the kiwi fruit at that moment is −46 feet per second per
second.
v0 ⫽ 0
v0
v(t)
h
v0
v(t)
⫺106 ft/s
s(t)
⫺200 ft/s
h
t⫽5
⫺46 ft/s 2
100 ft
t⫽3
h
h
t, seconds
a(t)
s(t)
t, seconds
199
26 ft
(a) Use the values of s (0) (note the kiwi initially has velocity zero), s(5), s (5), and s (5) given in the preceding description to find a formula for the height s(t) of
the kiwi fruit t seconds after being dropped from the
Xillian radio tower. Specifically, use these four data
points to solve for the coefficients of a cubic polynomial s(t) = at3 + bt2 + ct + d.
(a) If the height s(t) of the bowling ball t seconds after
being thrown is given by a quadratic polynomial
function, use s(3), s (3), and s (3) to find an equation
for s(t).
(b) Verify that the function s(t) you just found produces
the correct values for s (0), s(5), s (5), and s (5) in this
exercise.
(b) How high is the 20th-story window from which the
bowling ball was thrown?
(c) How high is the Xillian radio tower from which the
kiwi fruit was dropped?
(c) How fast was the bowling ball initially thrown?
(d) On Earth, acceleration due to gravity is given by a
constant −32 feet per second per second, that is,
on Earth we always have a(t) = −32 for falling objects. What is the function a(t) for acceleration due to
Xillian “gravity”? Is this acceleration constant?
What are the physical implications of gravity on
Xillian?
81. On Earth, a falling object has a downward acceleration of −32 feet per second per second due to gravity. Suppose an object falls from an initial height of
s0 feet, with an initial velocity of v0 feet per second. Use antiderivatives to show that the equations for
Proofs
84. Use the definition of the derivative to directly prove the
differentiation rules for constant and identity functions.
85. Use the h → 0 definition of the derivative to prove that
the power rule holds for positive integer powers.
86. Prove, in two ways, that the power rule holds for negative
integer powers:
(a) By using the z → x definition of the derivative.
(b) By using the h → 0 definition of the derivative.
87. Prove the difference rule in two ways:
(a) Using the definition of the derivative.
(b) Using the sum and constant multiple rules.
88. Use the definition of the derivative to prove the following
special case of the product rule:
d 2
(x f (x)) = 2 x f (x) + x 2 f (x).
dx
CONFIRMING PAGES
23:52
TKmaster2010
WHF00153/FREE087-Taalman
200
Chapter 2
November 21, 2012
Derivatives
89. Use the definition of the derivative to prove the quotient
rule. (Hint: At some point you will have to add and subtract
the quantity f (x)g(x) from the numerator.)
90. The following reciprocal rule tells us how to differentiate
the reciprocal of a function:
1
f (x)
d
=−
.
dx f (x)
( f (x))2
(a) Prove this antidifferentiation formula. You may assume that any two functions with the same derivative differ by a constant, as we will prove in
Section 3.2.
(b) What part of your argument from part (a) breaks
down when f (x) = x−1 ?
92. Consider the piecewise-defined function
Prove the reciprocal rule in two ways:
f (x) =
(a) By using the definition of the derivative.
(b) By using the quotient rule.
91. Consider the following formula for antidifferentiating
power functions: If f (x) = x k and k = −1, then f (x) =
1
x k + 1 + C for some constant C.
k+1
g(x), if x ≤ c
h(x), if x > c.
Prove that if g(x) and h(x) are continuous and differentiable at x = c, and if g(c) = h(c) and g (c) = h (c), then f
is differentiable at x = c.
Thinking Forward
For each function f that follows, find all of the x-values in the
domain of f for which f (x) = 0 and all values for which f (x)
does not exist. In later sections we will call these values the
critical points of f .
f (x) = x 3 − 2x
f (x) =
√
x−x
1
√
1+ x
f (x) =
x 2 (x − 1)
(x − 2)2
f (x) =
Taylor polynomials: In the exercises that follow, you will investigate the relationship between the coefficients of a polynomial
and its higher order derivatives. In later chapters this same
sort of idea will be used to locally approximate differentiable
functions with polynomials.
Prove that if f is any quadratic polynomial function
f (x) = ax 2 + bx + c, then the coefficients of f are completely determined by the values of f (x) and its derivatives at x = 0, as follows:
a=
f (0)
,
2
b = f (0),
and
c = f (0).
Prove that if f is any cubic polynomial function f (x) =
ax 3 + bx 2 + cx + d, then the coefficients of f are completely determined by the values of f (x) and its derivatives at x = 0, as follows:
a=
This means that every cubic polynomial function can
be written f (x) =
f (0) 3
f (0) 2
x +
x + f (0)x + f (0).
6
2
Suppose f is any cubic polynomial function
f (x) = ax 3 + bx 2 + cx + d. Prove that the coefficients a, b, c, and d of f can be expressed in terms
of the values of f (x) and its derivatives at the point
x = 2. (Hint: In other words, show that you can write the
coefficients a, b, c, and d in terms of f (2), f (2), f (2), and
f (2).)
Suppose f is a polynomial of degree n, and let k be
some integer with 0 ≤ k ≤ n. Prove that if f (x) is of
the form
f (x) = a n x n + a n−1 x n−1 + · · · + a k x k + · · · + a1 x + a 0 ,
In particular, this means that if f is any quadratic polynomial function, then f (x) =
f (0)
f (0)
, b=
, c = f (0),
6
2
and d = f (0).
f (0) 2
x + f (0)x + f (0).
2
then a k =
f (k) (0)
, where f (k) (x) is the kth derivative of
k!
f (x) and k! = k(k − 1) · · · (2)(1). (Hint: Find f (k) (x), and
use it to show that f (k) (0) = k!a k .)
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
2.4
201
The Chain Rule and Implicit Differentiation
THE CHAIN RULE AND IMPLICIT DIFFERENTIATION
The chain rule for differentiating compositions of functions
Implicit functions and implicit differentiation
Using implicit differentiation to prove derivative formulas
Differentiating Compositions of Functions
With the rules we have developed so far, we can differentiate any arithmetic combination
of functions whose components have known derivatives. There is, however, one other way
that functions can be combined: by composition. We still don’t know how to differentiate,
say, the composition y = (x 2 + 1) 1/2 . Is the derivative of a composition f ( g(x)) somehow
related to the derivatives of f and g?
Let’s consider a simple example involving rates of change. Suppose you own a small
factory that makes 30 widgets an hour and you make a profit of $10.00 for each widget
made. This means that you can make a profit of $300.00 an hour by making widgets at
your factory. We arrive at this answer by using the product:
dollars
dollars
widgets
300
30
.
= 10
hour
widget
hour
What does this have to do with the derivative of a composition? Let w(t) be the number
of widgets you have t hours after starting production, and let p(w) be the profit made from
producing w widgets. Then the composition p(w(t)) gives the profit made after t hours. We
dp
are interested in the rate of change of profit per hour, which is the derivative . The fact
that you make a profit of
dw
dt
dp
dw
dt
= 10 dollars per widget and the fact that your factory can make
= 30 widgets per hour are also statements about derivatives. Repeating our previous
calculation, we can relate the derivative of the composition p(w(t)) to the derivatives of p(w)
and w(t):
dp
dp dw
=
.
dt
dw dt
This example suggests that the derivative of a composition is the product of the derivatives of the component functions. That is in fact the case in general:
THEOREM 2.12
The Chain Rule
Suppose f (u(x)) is a composition of functions. Then for all values of x at which u is
differentiable at x and f is differentiable at u(x), the derivative of f with respect to x is
equal to the product of the derivative of f with respect to u and the derivative of u with
respect to x.
In Leibniz notation, we write this as
In “prime” notation, we write it as
df
df du
=
.
dx
du dx
( f ◦ u) (x) = f (u(x)) u (x).
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
202
Chapter 2
November 21, 2012
Derivatives
To use the chain rule, we must first recognize a function as a composition f (u(x)) and identify the “outside” and “inside” functions f and u. For example, the function y = (x 2 + 1) 1/2
can be thought of as a composition y = f (u(x)) with inside function u(x) = x 2 + 1 and outside function f (u) = u 1/2 . The chain rule says that we should differentiate the outside function f with respect to u and then multiply the result by the derivative of the inside function:
1
2
1
2
( f (u(x)) = f (u(x)) u (x) = (u(x))− 1/2 u (x) = (x 2 + 1)− 1/2 (2x).
For more examples of how to apply the chain rule, see Examples 1 and 2. Most of the
examples of the chain rule in this section will be fairly simple. We will do more complicated
examples in later sections after we have seen how to differentiate exponential, logarithmic,
and trigonometric functions.
The “prime” notation in Theorem 2.12 makes clear why we want u to be differentiable
at x and f to be differentiable at u(x). When it suits our purposes, we can also write ( f (u(x)))
for ( f ◦ u) (x). The Leibniz notation suggests why we call it the “chain” rule: We are taking
the derivative of a chain of functions by multiplying a chain of derivatives. If we had a longer
chain of functions, then we would have a longer chain of derivatives; for example, the
derivative of the composition f (u(v(x)) with respect to x is given by the chain of derivatives
df
df du dv
=
.
dx
du dv dx
The chain rule seems sensible if you consider its analog with difference quotients, since
f
f u
we can cancel u’s to justify the equation
=
. However, this cancellation does
x
u x
not automatically apply to the Leibniz notation, since the differentials df , dx, and du do not
represent numerical quantities. They cannot be cancelled just because the notation makes
it look tempting to do so. The proof of the chain rule requires more work than simply
“cancelling,” but except for a certain technical point, is not that difficult.
Proof. Our proof will start with the definition of the derivative for ( f ◦ g). After some algebra and
a limit rule, a change of variables will give us the result we desire. In the middle of the calculation
we will make a simplifying substitution:
( f ◦ g) (x) = lim
h→0
f ( g(x + h)) − f ( g(x))
h
= lim
h→0
=
f ( g(x + h)) − f ( g(x)) g(x + h) − g(x)
g(x + h) − g(x)
h
h→0
f ( g(x + h)) − f ( g(x))
g(x + h) − g(x)
k→0
k
lim
← derivative
lim
h→0
g(x + h) − g(x)
h
← algebra
← product rule for limits
f ( g(x) + k) − f ( g(x)) g (x)
= lim
← see below
= f ( g(x)) g (x)
← derivative
There is one technical point to consider in this proof. In the fourth step we applied the substitution
k = g(x + h) − g(x). Since g(x) is differentiable, it is also continuous; therefore as h → 0, we also
have k = g(x + h) − g(x) → 0. However, the preceding calculation assumes that k = g(x + h) − g(x)
is nonzero for small enough h, which in some situations may not be the case. In the case when this
happens because g(x + h) = g(x) as h→0, it is easy to show that ( f ◦ g) (x) and f ( g(x))g (x) are both
zero and therefore equal. There are other cases when this can happen, but handling the details for
those functions is beyond the scope of this course, so we will assume here that g(x) is nice enough
that we can avoid those cases.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
203
The Chain Rule and Implicit Differentiation
Implicit Differentiation
Consider the equation x 2 + y2 = 1 that describes the circle of radius 1 centered at the origin,
as shown next at the left. Clearly, this graph does not represent a function, since it does not
pass the vertical line test. However, locally, that is, in small pieces, the graph does define y
as a function of x. For example, the top half of the graph shown in the middle figure does
represent a function, as does the graph of the bottom half shown at the right.
x 2 + y2 = 1
y=
√
y = − 1 − x2
√
1 − x2
y
y
1
y
1
1
1
x
1
1
1
1
x
1
1
1
x
1
Although we cannot solve the equation x 2 + y2 = 1 for y and obtain a single well-defined
function, we can still think of the x-values as inputs and the y-values as outputs; the only
difference is that there may be more than one y-value for each x-value. In cases such as
these, we say that y is an implicit function of x.
Thinking locally, we can use our usual differentiation techniques to show that the unit
circle has horizontal
tangent lines at (0, 1) and at (0, −1). Looking at the top of the circle,
√
we have y = 1 − x 2 , with derivative
1
2
y = (1 − x 2 )− 1/2 (−2x) = √
−x
1 − x2
,
which is clearly zero√only when x = 0. Therefore the unit circle has a horizontal tangent
2
line
√ at the point (0, 1 − 0 ) = (0, 1). In a similar fashion we could use the equation y =
− 1 − x 2 for the bottom half of the circle to find the other horizontal tangent line.
We will not always be able to divide an implicit function into pieces given by actual
functions whose equations we know. That is, we will not always be able to “solve for y.”
However, given an equation that defines an implicit function, we can still find information
about slopes and derivatives simply by differentiating both sides of the equation with respect to x. This technique is known as implicit differentiation. The key to implementing
the technique will be applying the chain rule appropriately.
For example, we can differentiate both sides of the equation x 2 + y2 = 1 from the
previous example. Along the way we will have to remember that we are thinking of
y = y(x) as a function of x and apply the chain rule:
x 2 + y2 = 1
d 2
d
(x + ( y(x))2 ) = (1)
dx
dx
2x + 2y
← given equation
← differentiate both sides
dy
=0
dx
← power and chain rules
dy
−2x
x
=
=− .
dx
2y
y
← solve for
dy
dx
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
204
Chapter 2
November 21, 2012
Derivatives
The use of the chain rule in the differentiation step is crucial to these types of calculations.
Notice that the derivative of x 2 with respect to x is the familiar 2x, but the derivative of y2
dy
with respect to x uses the chain rule to arrive at 2y . Here y is thought of as an implicit
dx
function of x, so the expression y2 is really the composition of the implicit function y(x) with
the squaring function. That is why the chain rule must come into play when we differentiate
y2 with respect to x.
The fact that
dy
dx
x
y
= − means that for any point (a, b) that lies on the graph of the circle
a
x 2 + y2 = 1, the slope of the tangent line to the circle at (a, b) is given by − . For example,
b
2 √ 2
√ 1
3
1
3
is on the graph of the circle because
+
= 1 and the slope of
the point ,
2 2
2
2
√ −1/2
1
1
3
= − √ . Similarly, at the point − , −
the line tangent to the circle at that point is √
3/2
2
3
2
1
the slope of the tangent line is also − √ , as shown next at the left. At the point (0, 1) on
3
0
1
1
− and
0
the circle, the tangent line has slope − = 0 and is thus horizontal, and at the point (1, 0)
the tangent line has undefined slope
Slope at (a, b) given by −
y
slo
pe
slo
pe
1
Horizontal and vertical tangent lines
slope 0
1
y
1
3
1
2
1
2
a
b
x
undefined slope
3
2
thus is vertical, as shown at the right.
1
1
x
3
2
1
3
Examples and Explorations
EXAMPLE 1
Applying the chain rule one step at a time
4
x
, writing out all steps.
Use the chain rule to differentiate h(x) =
1 − 3x 2
SOLUTION
We can think of h(x) as a composition f (u(x)) with inside function u(x) =
side function f (u) = u4 . Applying the chain rule, we have
h (x) = f (u(x))u (x)
3 x
and out1 − 3x 2
← chain rule
= 4(u(x)) u (x)
3
x
u (x)
=4
1 − 3x 2
3 x
1(1 − 3x 2 ) − x(−6x)
=4
.
1 − 3x 2
(1 − 3x 2 )2
← derivative of outside function
← evaluate at inside function
← derivative of inside function
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
The Chain Rule and Implicit Differentiation
205
When you become comfortable using the chain rule, you won’t write out each step separately as we just did; instead you will simply think something like “the derivative of the
outside with the inside plugged in, times the derivative of the inside,” and write out the
derivative calculation in one step. Note that once again we have not simplified the derivative that we found. In this text we will not simplify such answers unless we have a specific
use for the derivative in question.
EXAMPLE 2
Determining which differentiation rule to apply first
For each function f , find the derivative f . What is the first differentiation rule that you must
use in each case?
√
√
x2 − 1
1 −2
3
(a) f (x) =
x−
(b) g(x) =
(c) h(x) = (x 2 −1)5 (3x−1)2
1
x
1−
x
SOLUTION
(a) The function f is at its outermost layer a composition of the form (u(x))−2 , where
√
1
u(x) = 3 x − . Therefore, the first rule to use is the chain rule. The power rule is
x
the second differentiation rule we must use, since we need to use the chain rule to even
begin differentiating the function. Before applying the chain rule, we use algebra to
rewrite the function f in a more convenient exponent form:
√
d
1 −2
d
3
= ((x 1/3 − x−1 )−2 )
x−
← algebra
dx
x
dx
= −2(x 1/3 − x−1 )−3 ·
= −2(x 1/3 − x−1 )−3
d 1/3
(x − x−1 )
dx
← chain, power rules
1 − 2/3
x
− (−1)x−2 .
3
← difference, power rules
(b) The function g(x) is at its outermost layer a quotient, so we will begin by using
the quotient rule. Again we write roots and fractions as exponents first, so that the
differentiation steps will be easier:
√
2
1/2 x2 − 1
d
d (x − 1)
=
← algebra
dx 1 − (1/x)
dx
1 − x−1
=
d
d
((x 2 − 1) 1/2 )(1 − x−1 ) − (x 2 − 1) 1/2 (1 − x−1 )
dx
dx
(1 − x−1 )2
← quotient rule
1 2
(x − 1)− 1/2 (2x)(1 − x−1 ) − (x 2 − 1) 1/2 (0 − (−1)x−2 )
2
. ← more rules
=
(1 − x−1 )2
(c) The function h(x) is at its outermost layer a product, so we begin by applying the product rule:
d
d
d
((x 2 −1)5 (3x−1)2 ) = ((x 2 −1)5 )(3x−1)2 + (x 2 −1)5 ((3x−1)2 )
dx
dx
dx
← product rule
= 5(x 2 −1)4 (2x)(3x−1)2 + (x 2 −1)5 (2)(3x−1)(3). ← more rules
EXAMPLE 3
Implicit differentiation and the chain rule
Each of the equations that follow defines y = y(x) as an implicit function of x. Use implicit
differentiation to find y = y (x).
(a)
3x 4 + 2y4 + x + y = 0
(b) x 2 y + xy2 = 1
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
206
Chapter 2
November 21, 2012
Derivatives
SOLUTION
(a) Before we begin, note that y = y(x) is an implicit function of x, so when we differentiate
any expression involving y with respect to x, we will need the chain rule. We will write
y and y to avoid any confusion with multiplication here, but each time we do so we
will think of the (implicit) functions y(x) and y (x). Differentiating both sides of the
given equation and then solving for y gives
d
d
(3x 4 + 2y4 + x + y) = (0)
dx
dx
← differentiate both sides
3(4x 3 ) + 2(4y3 )y + 1 + y = 0
3 ← note x and y are dealt with differently
← isolate y terms to one side
3
8y y + y = −12x − 1
y (8y3 + 1) = −12x 3 − 1
y =
← factor out y on the left
−12x 3 − 1
.
8y3 + 1
← divide to finish solving for y (b) This time we will need the product rule, since x 2 y is the product of the function x 2 and
the implicit function y, and xy2 is the product of the function x and the implicit function
y2 . For the latter product we will also require the chain rule. The overall process is much
the same as in the previous calculation:
d 2
d
(x y + xy2 ) = (1)
dx
dx
← differentiate both sides
(2xy + x 2 y ) + (1y2 + x(2yy )) = 0
2 ← product and chain rules
x y + 2xyy = −2xy − y
2
y (x 2 + 2xy) = −2xy − y2
y =
EXAMPLE 4
−2xy − y2
.
x 2 + 2xy
← isolate y terms to one side
← factor out y on the left
← divide to finish solving for y Finding values and tangent lines on the graph of an implicit function
Consider the equation y3 + xy + 2 = 0 that defines y as an implicit function of x.
(a) Show that y3 − 5y + 2 factors as ( y − 2)( y2 + 2y − 1).
(b) If x = −5, find all the possible values for y.
(c) Find the slope of the line tangent to y3 + xy + 2 = 0 at the point (−5, 2).
SOLUTION
(a) We could factor y3 − 5y + 2 by using synthetic division, which you may have seen in
a previous course, but since the factorization is already given in the problem, all we
need to do is multiply out and check:
( y − 2)( y2 + 2y − 1) = y3 + 2y2 − y − 2y2 − 4y + 2 = y3 − 5y + 2.
(b) Given the factorization from part (a), we substitute x = −5 into the equation
y3 + xy + 2 = 0 and solve for all corresponding values of y:
y3 − 5y + 2 = 0
← the equation with x = −5
2
( y − 2)( y + 2y − 1) = 0
← factorization in part (a)
√
√
← quadratic formula
y = 2, y = −1 + 2, or y = −1 − 2.
√
√
Thus (−5, 2), (−5, −1 + 2 ), and (−5, −1 − 2 ) are points on the graph of
y3 + xy + 2 = 0.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
(c) To find the slope of the tangent line we must first calculate
tiation and the chain rule, we have
y3 + xy + 2 = 0
3y2
dy
dy
+ (1)( y) + (x)
dx
dx
3y2
dy
. Using implicit differendx
← the given equation
d
d
(( y(x))3 + x · y(x) + 2) = (0)
dx
dx
207
The Chain Rule and Implicit Differentiation
← differentiate both sides
+0=0
← chain and product rules
dy
dy
+ x = −y
dx
dx
← start solving for
dy
(3y2 + x) = −y
dx
−y
dy
= 2
.
dx
3y + x
dy
dx
← algebra
← algebra
This means that if a point (a, b) is on the graph of y3 + xy + 2 = 0, then the line tangent
to the graph of y3 + xy + 2 = 0 at (a, b) has slope
−b
dy .
x=a = 2
dx y = b
3b + a
Therefore the slope of the tangent line at the point (a, b) = (−5, 2) is
−2
−2
−2
dy =
=
.
x = −5 =
dx y = 2
3(2)2 + (−5)
12 − 5
7
CHECKING
THE ANSWER
We can use a graphing utility to sketch the graph of the implicit function from the previous
example, as shown next at the left. Notice that for x = −5 there are three corresponding
y-values, as we showed in part (b). At each of these points (−5, y) there is a tangent line
to the graph. From the graph at the right it does appear (taking axes scales into account)
2
that the slope of the tangent line at (−5, 2) could be approximately − .
7
y3 + xy + 2 = 0 and the points at x = −5
Slope at (−5, 2) is −
y
10
EXAMPLE 5
y
4
4
2
2
5
2
7
5
10
x
10
5
5
2
2
4
4
10
x
Using implicit differentiation to prove the general power rule
Although we have been using the power rule for general rational powers, we have in fact
proved it only for integer powers.
(a) Use implicit differentiation and the power rule for integer powers to prove that
d
5
(x 5/3 ) = x 2/3 .
dx
3
(b) Use implicit differentiation and the power rule for integer powers to prove the power
rule for rational powers.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
208
Chapter 2
November 21, 2012
Derivatives
SOLUTION
(a) If y = x 5/3 , then y3 = x 5 . With this equation we can use implicit differentiation and
the power rule for integer powers to find y :
y3 = x 5
← since y = x 5/3
d
d
(( y(x))3 ) = (x 5 )
dx
dx
← differentiate both sides
3y2 y = 5x 4
y =
← chain rule and integer power rule
5x 4
5x 4
← solve for y , substitute y = x 5/3
=
3y2
3(x 5/3 )2
5
5
y = x4−(10/3) = x 2/3 .
3
3
(b) Suppose y = x p/q , where
p
q
← algebra
is a rational number. Generalizing the method used in
part (a), we use implicit differentiation on the equation y q = x p to solve for y :
yq = xp
← since y = x p/q
d
d
(( y(x))q ) = (x p )
dx
dx
← differentiate both sides
qy q−1 y = px p−1
← chain rule, integer power rule
px p−1
px p−1
y = q−1 =
qy
q(x p/q )q−1
← solve for y , substitute y = x p/q
px p−1
p
= x((p−1)−( p(q − 1)/q))
q
qx p(q − 1)/q
p
y = x p/q − 1 .
q
y =
TEST YOUR
? UNDERSTANDING
← algebra
← even more algebra
What is the difference between f ( g(x)) and ( f ( g(x))) ?
The chain rule formula f (u(x))u (x) is not exactly the same as the product of the deriva-
tives of f and u. What is the difference?
What is the difference between saying that y is a function of x and saying that y is an
implicit function of x?
Suppose y is an implicit function of x. Would it be correct or incorrect to say that
d
( y3 )
dx
= 3y2 , and why?
If f is an invertible function with inverse f −1 , then what is the relationship between the
derivatives of f and f −1 ?
EXERCISES 2.4
Thinking Back
Differentiation review: Without using the chain rule, find the
derivatives of each of the functions f that follow. Some algebra may be required before differentiating.
1
1 2
+ 2
f (x) =
f (x) = (3x + 1)4
x
x
√
f (x) = (x + 1)2 x
f (x) =
(x + 1)2
√
x
Decomposing functions: For each function k that follows, find
functions f , g, and h so that k = f ◦ g ◦ h. There may be more
than one possible answer.
k(x) = (x 2 + 1)3
k(x) =
1
3x + 1
k(x) =
1 + (x − 2)2
k(x) = √
CONFIRMING PAGES
1
3x + 1
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
209
The Chain Rule and Implicit Differentiation
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The chain rule is used to differentiate
compositions of functions.
(b) True or False: If f and g are differentiable functions,
then the derivative of f ◦ g is equal to the derivative
of g ◦ f .
(c) True or False: If f and g are differentiable functions,
then
d
( f ( g(x))) = f (x)g (x).
dx
(d) True or False: If u and v are differentiable functions,
then
d
(u(v(x))) = u (v (x)).
dx
√
7. Differentiate f (x) = (3x + x )2 in three ways. When
you have completed all three parts, show that your three
answers are the same:
(a) with the chain rule
(b) with the product rule but not the chain rule
(c) without the chain or product rules
4 3
x −2
8. Differentiate f (x) =
in three ways. When you
√
x
have completed all three parts, show that your three answers are the same:
(a) with the chain rule
(b) with the quotient rule but not the chain rule
(c) without the chain or quotient rules
Suppose g, h, and j are differentiable functions with the values
for the function and derivative given in the following table:
(e) True or False: If h and k are differentiable functions,
then
d
(k(h(x))) = k (h(x))h (x).
dx
(f) True or False: If y is an implicit function of x, then
there can be more than one y-value corresponding
to a given x-value.
(g) True or False: The graph of an implicit function can
have vertical tangent lines.
(h) True or False: If y is an implicit function of x and
dy = 0, then the graph of the implicit function
dx x=2
has a horizontal tangent line at (2, 0).
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Three functions that we could not have differentiated
before learning the chain rule, even after algebraic
simplification.
(b) An equation that defines y as an implicit function
of x, but not as a function of x.
(c) The graph of an implicit function with three horizontal tangent lines and two vertical tangent lines.
3. State the chain rule for differentiating a composition
g(h(x)) of two functions expressed (a) in “prime” notation
and (b) in Leibniz notation.
4. In the text we noted that if f (u(v(x))) was a composition
of three functions, then its derivative is
df
df du dv
=
.
dx
du dv dx
Write this rule in “prime” notation.
5. Write down a rule for differentiating a composition
f (u(v(w(x)))) of four functions (a) in “prime” notation and
(b) in Leibniz notation.
√
u2 + 3u5
6. Suppose u(x) = 3x 2 + 1 and f (u) =
. Use the
d
1−u
chain rule to find
( f (u(x))) without first finding the
dx
formula for f (u(x)).
x
−3
−2
−1
0
1
2
3
g(x)
0
1
3
2
0
−2
−3
h(x)
3
2
0
3
−1
−2
0
j(x)
1
3
1
0
−2
−3
1
g (x)
1
2
−1
−2
−2
−1
0
h (x)
0
−3
−2
3
−2
0
1
j (x)
2
0
−2
−2
−1
2
2
Use the table to calculate the values of the derivatives listed
in Exercises 9–16.
9. If f (x) = g(h(x)), find f (3).
10. If f (x) = h( g(x)), find f (3).
11. If f (x) = ( g(x))3 , find f (−2).
12. If f (x) = g(x 3 − 6), find f (2).
13. If f (x) = h( g(j(x))), find f (1).
14. If f (x) = j(2x), find f (−1).
15. If f (x) = h( g(x)j(x)), find f (0).
16. If f (x) = h(h(h(x))), find f (1)
17. If y is a function of x, then how is the chain rule involved
in differentiating y3 with respect to x, and why?
18. Show that, for any integers p and q (with q = 0),
( p − 1) −
p
p(q − 1)
= − 1.
q
q
What does this equation have to do with the current
section?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
210
Chapter 2
November 21, 2012
Derivatives
19. Match the two graphs shown here to the equations
(x + 1)( y2 + y − 1) = 1 and xy2 + y = 1. Explain your
choices.
y
y
y
3
3
2
2
1
1
2 1
1
20. Match the two graphs shown here to the equations
xy2 + x = 1 and 1 + x + xy2 = 0. Explain your choices.
1
2
3
4
x
4 3 2 1
1
2
2
3
3
y
3
3
2
2
1
1
2
x
1
x
1
1
1
2
2
3
3
1
x
Skills
Find the derivatives of the functions in Exercises 21–46. Keep
in mind that it may be convenient to do some preliminary
algebra before differentiating.
√
1
22. f (x) = ( x + 4)5
21. f (x) = 3
x +1
√
23. f (x) = x(3x 2 + 1)9
24. f (x) = x 2 + 1
3x + 1
26. f (x) = √
x2 + 1
√
(1 + x )2
28. f (x) = 2
3x − 4x + 1
1
25. f (x) = √
2
x +1
√
27. f (x) = (x x + 1)−2
1
− 3x 2
29. f (x) = x
1
x5 − √
x
(x + 1)(3x − 4)
√
x 3 − 27
√
32. f (x) = 2 − 3x + 1
30. f (x) =
31. f (x) = (x 1/3 − 2x)−1
33. f (x) = x− 1/2 (x 2 − 1)3 34. f (x) = (x− 1/2 (x 2 − 1))3
Calculate each of the derivatives or derivative values in
Exercises 47–52.
√
d2
((x x + 1)−2 )
47.
2
dx
⎛
⎞
1
2
2
−
3x
d ⎜x
⎟
48.
⎝
⎠
dx 2 x 5 − √1
d2
49.
dx 2
x
√
x− 3/2 x x=2
3
d2
((5x 4 − 3x 2 )7 (2x 3 + 1))
dx 2
d3
(x(3x 2 + 1)9 )|x=0
51.
dx 3
x2 + 1
d
52.
dx (x 2 + 4)(3x − 2) 50.
x=−1
Suppose that r is an independent variable, s is a function of r,
and q is a constant. Calculate the derivatives in Exercises 53–
58. Your answers may involve r, s, q, or their derivatives.
3x − 4(2x + 1)6
√
x 2 − 3 x + 5x 9
36. f (x) =
√
x−1
35. f (x) =
53.
d 3
(s )
dr
d
(sr 2 )
dr
54.
d 3
(r )
dr
d
(rs2 )
dr
55.
d 3
(q )
dr
d
(qs2 )
dr
37. f (x) = (5(3x 4 − 1)3 + 3x − 1)100
56.
38. f (x) = (1 − 4x) (3x + 1)
For each of the equations in Exercises 59–62, y is defined as
an implicit function of x. Solve for y, and use what you find to
sketch a graph of the equation.
1 2
60. (x − 1)2 + ( y + 2)2 = 4
x + y2 = 9
59.
4
62. 4y2 − x 2 + 25 = 0
61. x 2 − 3y2 = 16
2
2
9
39. f (x) = 3((x 2 + 1)8 − 7x)− 2/3
40. f (x) =
(3x + 1)(x 4 − 3)4
(x + 5)−2 (1 + x 2 )5
41. f (x) = (5x 4 − 3x 2 )7 (2x 3 + 1)
√
√
42. f (x) = x 3x 2 + 1 3 2x + 5
43. f (x) = ((2x + 1)−5 − 1)−9
√
44. f (x) = x(x 2 )( x )(x 2/3 )
√
45. f (x) = (x 4 − 3 − 4x )8 + 5x
√
46. f (x) = 1 + 1 + 1 − 2x
57.
58.
In Exercises 63–68, find a function that has the given derivative and value. In each case you can find the answer with an
educated guess-and-check process while thinking about the
chain rule.
63. f (x) = 5(x 2 + 1)4 (2x), f (0) = 1
64. f (x) = 5(x 2 + 1)4 (2x), f (1) = 25
65. f (x) = x(x 2 + 1)4 , f (0) = 0
66. f (x) = −6x(x 2 + 1)4 , f (0) = 1
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.4
√
3x + 1, f (0) = 1
x2
, f (1) = 2
68. f (x) = √
x3 + 1
Each of the equations in Exercises 69–80 defines y as an
implicit function of x. Use implicit differentiation (without
67. f (x) =
solving for y first) to find
dy
.
dx
69. 4x 2 − y2 = 9
70. x 2 + y2 = 4
71. xy2 + 3x 2 = 4
72. y6 − 3x + 4 = 0
73. (3x + 1)( y2 − y + 6) = 0
75.
3y − 1 = 5xy
74. x 2 y − y2 x = x 2 + 3
y2 + 1
=x
77.
3y − 1
1
x2
1
− =
79.
y
x
y+1
78. 3y = 5x + 3 y − 2
76. (3y2 + 5xy − 2)4 = 2
2
1
x+1
=
80. 2
y −3
xy
In Exercises 81–84, use implicit differentiation to algebraically
find each quantity or location related to the given implicit
function.
81. Consider the circle of radius 1 centered at the origin, that
is, the solutions of the equation x 2 + y2 = 1.
(a) Find all points on the graph with an x-coordinate of
x=
1
, and then find the slope of the tangent line at
2
each of these points.
(b) Find √
all points on the graph with a y-coordinate of
2
y=
, and then find the slope of the tangent line
2
at each of these points.
(c) Find all points on the graph where the tangent line is
vertical.
(d) Find all points on the graph where the tangent line
has a slope of −1.
The Chain Rule and Implicit Differentiation
211
82. Consider the graph of the solutions of the equation
4y2 − x 2 + 2x = 2.
(a) Find all points on the graph with an x-coordinate of
x = 3, and then find the slope of the tangent line at
each of these points.
(b) Find all points on the graph with a y-coordinate of
y = 3, and then find the slope of the tangent line at
each of these points.
(c) Find all points where the graph has a horizontal
tangent line.
(d) Find all points where the graph has a vertical tangent
line.
83. Consider the graph of the solutions of the equation
y3 + xy + 2 = 0.
(a) Find all points on the graph with an x-coordinate of
x = 1, and then find the slope of the tangent line at
each of these points.
(b) Find all points on the graph with a y-coordinate of
y = 1, and then find the slope of the tangent line at
each of these points.
(c) Find all points where the graph has a horizontal tangent line.
(d) Find all points where the graph has a vertical tangent
line.
84. Consider the graph of the solutions of the equation
y3 − 3y − x = 1.
(a) Find all points on the graph with an x-coordinate of
x = −1, and then find the slope of the tangent line at
each of these points.
(b) Find all points on the graph with a y-coordinate of
y = 2, and then find the slope of the tangent line at
each of these points.
(c) Find all points where the graph has a horizontal
tangent line.
(d) Find all points where the graph has a vertical tangent
line.
Applications
85. Linda can sell 12 magazine subscriptions per week and
makes $4.00 for each magazine subscription she sells.
Obviously this means that Linda will make (12)($4.00) =
$48.00 per week from magazine subscriptions. Explain
this result mathematically, using mathematical notation
and the chain rule.
86. If you drop a pebble into a large lake, you will cause a circle of ripples to expand outward. The area A = A(t) and
radius r = r(t) are clearly functions of t (they change over
time) and are related by the formula A = πr 2 .
Area A A(t)
(a) If r is measured in inches and t is measured in
seconds, what are the units of
dA
? What are the units
dt
dA
?
dr
dA
and explain the practical meaning of
(b) Find
dr
dA
.
dr r=2
dA
(c) Find
and explain the practical meaning of
dt
dA
.
dt r=2
of
87. The area of a circle can be written in terms of its radius as
A = π r 2 , where both A and r are functions of time. Suppose a circular area from a spotlight on a stage floor is
slowly expanding.
r r(t)
t, seconds
dA
and explain its meaning in practical terms.
dr
dA
depend on how fast the radius of
(b) Does the rate
dr
(a) Find
the circle is increasing? Does it depend on the radius
of the circle? Why or why not?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
212
Chapter 2
November 21, 2012
Derivatives
dA
and explain its meaning in practical terms.
dt
dA
depend on how fast the radius of
(d) Does the rate
dt
(c) Find
the circle is increasing? Does it depend on the radius
of the circle?
(e) If the radius of the circle of light is increasing at a constant rate of 2 inches per second, how fast is the area
of the circle of light increasing at the moment that the
spotlight has a radius of 24 inches?
Proofs
d
88. Use the chain rule twice to prove that ( f (u(v(x)))) =
dx
f (u(v(x)))u (v(x))v (x).
89. In Exercise 89 of the Section 2.3 you used the definition of
derivative to prove the quotient rule. Prove it now another
f
way: by writing a quotient as a product and applying the
g
product, power, and chain rules. Point out where you use
each rule.
d
90. Use implicit differentiation and the fact that (x 4 ) = 4x 3
dx
d
to prove that (x−4 ) = −4x−5 .
dx
d 3
(x ) = 3x 2
dx
d
d
3
and (x 5 ) = 5x 4 to prove that (x 3/5 ) = x− 2/5 .
dx
dx
5
91. Use implicit differentiation and the fact that
92. Use implicit differentiation and the power rule for integer powers (not the general power rule) to prove that
d 2/3
2
(x ) = x− 1/3 .
dx
3
93. Use implicit differentiation, the product rule, and the
power rule for positive integer powers to prove the power
rule for negative integer powers.
94. Use implicit differentiation and the power rule for integer
powers to prove the power rule for rational powers.
Thinking Forward
Finding critical points: For each of the following functions f ,
find all of the x-values for which f (x) = 0 and all of the
x-values for which f (x) does not exist.
√
f (x) = x 3 3x + 1
f (x) = (1 − x 4 )7
f (x) = (x 2 + 3)(x − 2) 3/2
f (x) = 3x x +
√
x
f (x) = √
x x−1
2.5
1
x
√
Finding antiderivatives by undoing the chain rule: For each function f that follows, find a function F with the property that
F (x) = f (x). You may have to guess and check to find such a
function.
√
√
f (x) = x 1 + x 2
f (x) = x 2 1 + x 3
f (x) =
1
(2 − 5x)3
f (x) = √
1
1 + 3x
f (x) = (x x + 1 )−2
DERIVATIVES OF EXPONENTIAL AND LOGARITHMIC
FUNCTIONS
Formulas for differentiating exponential and logarithmic functions
Rates of growth of exponential functions
The method of logarithmic differentiation
Derivatives of Exponential Functions
The power rule tells us that the derivative of x k is kx k−1 . This rule works only for power
functions, where the base is the variable x and the power is a constant k; it does not tell
us how to differentiate an exponential function like 2 x or e 3x , where the variable is in the
exponent. To determine such derivatives we must return to the definition of the derivative:
b x+h − b x
b xb h − b x
b x (b h − 1)
d x
(b ) = lim
= lim
= lim
= bx
dx
h
h
h
h→0
h→0
h→0
bh − 1
.
h
h→0
lim
Although we have simplified as much as possible, this is a limit that we do not yet know
how to calculate.
CONFIRMING PAGES
22:0
WHF00153/FREE087-Taalman
November 21, 2012
2.5
213
Derivatives of Exponential and Logarithmic Functions
In one special case we do already know how to evaluate this limit. In Theorem 1.26
we saw that e is the unique number such that lim
h→0
eh − 1
h
= 1. Therefore when b = e, the
preceding calculation looks like this:
e x+h − e x
d x
(e ) = lim
= · · · = ex
dx
h
h→0
eh − 1
h
h→0
= e x (1) = e x .
lim
We have just shown that the function f (x) = e x is its own derivative! This is in fact the exact
reason that we defined the number e the way that we did. As illustrated here, y = e x is its
own associated slope function:
6
6
6
5
5
4
4
e1
p
slo
2
1
2
x
2
pe
5
4
3
3
2
2
height e
1
height 1
1
e
7
3
1
y
7
1
1
2
48
y
7
e 4.
y
2
Slope and height of y = e x
both equal e 1.5 ≈ 4.48 at x = 1.5
Slope and height of y = e x
both equal e ≈ 2.18 at x = 1
slop
Slope and height of y = e x
both equal 1 at x = 0
slo
TKmaster2010
x
height
4.48
1
2
1
1
2
x
Other exponential functions of the form e kx or b x have graphs similar to the graph of e x ,
but only e x is scaled in exactly the right way to be its own derivative. With the chain rule,
we can use the derivative of e x to find the derivatives of general exponential functions:
THEOREM 2.13
Derivatives of Exponential Functions
For any constant k, any constant b > 0 with b = 1, and all x ∈ R,
(a)
d x
(e ) = e x
dx
(b)
d x
(b ) = ( ln b) b x
dx
Proof. The previous discussion proves that
(e ln b ) x and then apply the chain rule:
d x
d
(b ) = ((e ln b ) x )
dx
dx
=
d ( ln b) x
(e
)
dx
= e ( ln b)x
d
(( ln b)x)
dx
(c)
d kx
(e ) = ke kx
dx
d x
(e ) = e x . To prove the second rule we rewrite b x as
dx
← b = e ln b
← algebra of exponents
← chain rule and derivative of e x
= (e ln b ) x ( ln b)
← algebra and derivative of linear function
= ( ln b) b x .
← algebra
The proof for e kx is similar, with a simpler application of the chain rule, and is left for
Exercise 69.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
214
Chapter 2
November 21, 2012
Derivatives
By combining our new differentiation rules for exponential functions with the chain
rule, we obtain the following rules for exponential compositions:
d u(x)
(e ) = e u(x) u (x),
dx
d u(x)
(b ) = ( ln b)bu(x) u (x).
dx
For example, we have
2
2
d x 2 +1 d
e
= e x +1 · (x 2 + 1) = e x +1 (2x),
dx
dx
2
2
d x 2 +1 d
2
= ( ln 2)2 x +1 · (x 2 + 1) = ( ln 2)2 x +1 (2x).
dx
dx
Exponential Functions Grow Proportionally to Themselves
Notice that all exponential functions have the property that their derivatives are constant
multiples of the original function; for example,
d 2x
(e )
dx
= 2e 2x is just 2 times the original
function e 2x . This means that the rate of change of an exponential function f is proportional
to the function f itself. In fact, the converse is also true, although we will not have the tools
to prove it until Section 7.5. Together, both proportionalities give us the following two-sided
statement, which will prove useful in many application problems:
THEOREM 2.14
Rates of Change and Exponential Functions
f (x) = kf (x) for some constant k if and only if f is an exponential function of the form
f (x) = Ae kx .
Recall that one quantity y is proportional to another quantity x if y is a constant multiple
of x. Therefore the preceding theorem gives us a nice characterization of exponential functions: All exponential functions f , and only exponential functions, have the property that f is proportional to f .
If a word problem states that the rate of change of a function is constant, we immediately
know that the function is linear. Similarly, by Theorem 2.14, if a word problem states that the
rate of change of a function is proportional to the function itself, we immediately know that
that function is exponential. Then it is just a question of finding values A and k to determine
the function f (x) = Ae kx that models the situation. Exponential functions often arise when
modeling populations, for example, since a larger population can produce greater numbers
of offspring than a smaller population and can therefore grow at a faster rate.
Derivatives of Logarithmic Functions
Up to this point, the derivatives that we have seen are similar to their original functions.
For example, derivatives of power functions are power functions, derivatives of polynomials
are polynomials, and derivatives of exponential functions are exponential. Now something
surprising happens: Derivatives of logarithmic functions are not logarithmic. Even more
surprisingly, logarithmic functions are transcendental but their derivatives are algebraic!
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.5
THEOREM 2.15
215
Derivatives of Exponential and Logarithmic Functions
Derivatives of Logarithmic Functions
For any constant b > 0 with b = 1, and all appropriate values of x,
(a)
d
1
(logb x) =
dx
( ln b)x
(b)
d
1
( ln x) =
dx
x
d
1
( ln |x|) =
dx
x
(c)
The second rule is a special case of the first, with b = e. Because ln x has domain (0, ∞),
d
1
1
when we say that ( ln x) = , we are also restricting to the domain (0, ∞). In the third
dx
x
x
rule we generalize the second rule so that we are considering the full domain (−∞, 0) ∪
1
x
(0, ∞) of .
Proof. We will prove the second rule and leave the proofs of the first and third rules to Exercises 72
and 73, respectively. Since y = ln x and y = e x are inverses, e ln x = x for all x in the domain (0, ∞)
of y = ln x. Differentiating both sides of the equation gives
e ln x = x
e ln x
← property of inverses
d ln x
d
(e ) = (x)
dx
dx
← differentiate both sides
d
( ln x) = 1
dx
← exponential and chain rules
d
1
( ln x) = ln x
dx
e
← solve for
d
1
( ln x) = .
dx
x
← since e ln x = x
d
( ln x)
dx
The two graphs that follow illustrate that the associated slope function for ln |x| is the
1
function . For negative values of x, as we move from left to right, the slopes of ln |x| are
x
negative with larger and larger magnitude while the heights of
1
x
behave the same way. For
positive values of x, as we move from left to right, the slopes of ln x are positive but getting
1
smaller while the heights of do the same.
x
Heights of f (x) =
Slopes of f (x) = ln |x|
y
y
4
3
2
2
1
4 3 2 1
1
1
2
3
1
x
1
2
3
4
x
4
2
2
4
x
2
4
The rule for differentiating ln |x| will come in handy for calculating certain derivatives.
Functions that involve many products or quotients or that have variables in both a base
and an exponent can be difficult to differentiate. One strategy for differentiating such
CONFIRMING PAGES
22:0
WHF00153/FREE087-Taalman
216
Chapter 2
November 21, 2012
Derivatives
functions is to apply ln |x| to both sides of the equation y = f (x) and then differentiate both sides. This process is called logarithmic differentiation and is illustrated in
Examples 4 and 5.
By combining our new differentiation rules for logarithmic functions with the chain
rule, we obtain the following rules for logarithmic compositions:
1
u (x)
d
( ln u(x)) =
· u (x) =
,
dx
u(x)
u(x)
1
u (x)
d
(log b u(x)) =
· u (x) =
.
dx
( ln b)u(x)
( ln b)u(x)
For example, we have
1
2x
d
( ln(x 2 + 1)) = 2
· 2x = 2
,
dx
x +1
x +1
1
2x
d
(log b (x 2 + 1)) =
· 2x =
.
dx
( ln b)(x 2 + 1)
( ln b)(x 2 + 1)
Derivatives of Inverse Functions*
We can generalize the technique used in the proof of Theorem 2.15 to obtain a formula for
the derivative of the inverse of any function whose derivative we already know. If f is an
invertible function, then the graph of its inverse y = f −1 (x) can be obtained by reflecting
the graph of y = f (x) over the line y = x. This reflection yields the reciprocals of all slopes
on the graph, as shown in the following figure:
The slope of f −1 (x) at (b, a) is the
reciprocal of the slope of f (x) at (a, b)
pe
f (x)
r
y
slo
TKmaster2010
b
1
pe r
slo
f 1 (x)
a
a
b
x
From this example we might expect that the derivatives of f and f −1 would have some
sort of reciprocal relationship. In particular, we would expect that if f (a) = b and thus
1
a = f −1 (b), then ( f −1 ) (b) = . Using implicit differentiation and the chain rule, we
f (a)
can show that this is indeed the case in general:
THEOREM 2.16
Derivatives of Inverse Functions
If f and f −1 are inverse functions and are both differentiable, then, for all appropriate
values of x,
1
( f −1 ) (x) = −1
.
f ( f (x))
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.5
217
Derivatives of Exponential and Logarithmic Functions
For example, if f (x) = x 3 , then its inverse is the function f −1 (x) = x 1/3 , since y = x 3 if and
only if y 1/3 = x. By Theorem 2.16, the derivative of f −1 (x) = x 1/3 must be
( f −1 ) (x) =
1
f
(
f
−1 (x))
1
=
3( f
−1 (x))2
=
1
3(x 1/3 )2
1
1
= x− 2/3 ,
2/3
3x
3
=
just as we would expect from the power rule.
Proof. Since f and f −1 are inverses, we know that their composition in either order is the identity
function. Starting from this fact and applying implicit differentiation, we have
f ( f −1 (x)) = x
← definition of inverses
d
d
( f ( f −1 (x))) = (x)
dx
dx
f ( f −1 (x))
← differentiate both sides
d
( f −1 (x)) = 1
dx
← chain rule
d
1
( f −1 (x)) = −1
.
dx
f ( f (x))
← solve for
d
( f −1 (x))
dx
This result holds whenever x is in the domain of f −1 and f is differentiable at f −1 (x).
Examples and Explorations
EXAMPLE 1
Differentiating combinations of exponential and logarithmic functions
Find the derivatives of each of the following functions:
2
(a) f (x) = e x ln x
(b) f (x) =
7e 3x − x 2 2 x
log5 x
(c) f (x) = ln
2
x −1 2
1 − 2x
SOLUTION
(a) Using the product and chain rules, we have
2
2
2
d x2
d x2
d
e ln x =
e
· ln x + e x · ( ln x) = 2xe x ln x + e x
dx
dx
dx
1
.
x
(b) The function f (x) is a quotient, so we begin by applying the quotient rule:
f (x) =
=
=
d
d
(7e 3x − x 2 2 x ) · log5 x − (7e 3x − x 2 2 x ) · (log5 x)
dx
dx
21e 3x −
(log5 x)2
d 2 x
(x 2 ) (log5 x) − (7e 3x − x 2 2 x )
dx
1
( ln 5)x
← quotient rule
(log5 x)2
← other rules
(21e 3x − (2x2 x + x 2 ( ln 2)2 x ))(log5 x) − (7e 3x − x 2 2 x )
1
( ln 5)x
(log5 x)2
← product rule
In the preceding calculation, we saved the product rule calculation for the last step. For
lengthy derivative problems it is often helpful to postpone some calculations to later
steps.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
218
Chapter 2
November 21, 2012
Derivatives
(c) There are a number of ways we could find f (x). One way is to jump right in and start
differentiating; although it is possible to do this all in one step, we’ll do the calculation
in a few separate steps to be as clear as possible:
2
2
x −1 2
1
x −1 2
d
d
·
ln
= ← chain rule
2 dx
dx
1 − 2x
1 − 2x
x2 − 1
1 − 2x
−1
= 2 (2)
2
1
−
2x
x −1
1
1 − 2x
= 1
x2 − 1
1 − 2x
2 (2)
x2
x2 − 1
1 − 2x
1
2
d x −1
·
dx 1 − 2x
1 ← chain rule
(2x)(1 − 2x) − (x 2 − 1)(−2)
.
(1 − 2x)2
← quotient rule
The first step used the chain rule and the logarithmic rule, the second step used the
chain rule and the power rule, and the last step used the quotient rule.
If we do a little bit of algebra first, the differentiation step becomes much easier:
2
2
x −1 2
x −1
d
d
← algebra
ln
=
2 ln
dx
dx
1 − 2x
1 − 2x
=
d
(2( ln(x 2 − 1) − ln(1 − 2x)))
dx
=2
1
1
(2x) −
x2 − 1
1 − 2x
(−2) .
← algebra
← differentiate
In the preceding calculation, the first two steps were algebra; only the final step involved differentiation. Although the two answers obtained look very different, they
are in fact the same. (As an exercise, use algebra to show this.)
EXAMPLE 2
Differentiating a piecewise-defined function
Find the derivative of the piecewise-defined function f (x) =
e x,
e −x ,
if x < 0
if x ≥ 0.
SOLUTION
From the differentiation rules we have developed, we know that
−e −x . These are the expressions for the derivative of f
d x
(e )
dx
= e x and
d −x
(e )
dx
=
when x < 0 and when x > 0, respectively. It remains only to determine what happens at the breakpoint x = 0.
If g(x) = e x and h(x) = e −x , then we must first check that g(0) = h(0); since e0 = 1 and
= 1, this is true. It follows that the function f is continuous at x = 0. Second, we must
check that g (0) = h (0). Here g (0) = e0 = 1 but h (0) = −e −0 = −1, so although f is
continuous at x = 0, it is not differentiable. The derivative of f is therefore
⎧
e x , if x < 0
⎨
f (x) = undefined, if x = 0
⎩
−e −x , if x > 0.
e −0
EXAMPLE 3
A real-world problem that can be modeled with an exponential function
Suppose a population of wombats on a small island is growing at a rate proportional to the
number of wombats on the island. If there were 12 wombats on the island in 1990 and 37
wombats on the island in 1998, how many wombats were on the island in the year 2010?
SOLUTION
We are given that the rate of change W (t) of the population of wombats is proportional
to the population W(t) of wombats at time t; in other words, W (t) = kW(t) for some real
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
2.5
Derivatives of Exponential and Logarithmic Functions
219
number k. By Theorem 2.14, this means that the population W(t) of wombats on the island
must be an exponential function W(t) = Ae kt . We now only need to find constants A and
k that match the information given in the problem.
If we let t = 0 represent 1990 (so that t = 8 will represent 1998), then from the information given in the problem, we have W(0) = 12 and W(8) = 37. We will use these two
data points to find A and k. Using the first data point, we have
W(0) = 12 =⇒ Ae k(0) = 12 =⇒ Ae0 = 12 =⇒ A(1) = 12 =⇒ A = 12,
so W(t) = 12e kt for some k. Using the second data point, we can now solve for the value
of k:
37
ln
37
37
12
k(8)
8k
W(8) = 37 =⇒ 12e
=⇒ k =
= 37 =⇒ e =
=⇒ 8k = ln
≈ 0.14.
12
12
8
Thus W(t) = 12e 0.14t . Using this function, we can now easily calculate the number of
wombats that were on the island in the year 2010. Since 2010 is 20 years after 1990, we
need to find
W(20) = 12e 0.14(20) ≈ 197.34.
In the year 2010, there were approximately 197 wombats on the island, assuming of course
that we cannot have “parts” of wombats.
EXAMPLE 4
Using logarithmic differentiation when products and quotients are involved
Use logarithmic differentiation to calculate the derivative of the function
√
x (x 2 − 1)5
.
f (x) =
(x + 2)(x − 4)3
SOLUTION
Obviously it would take a great deal of work to differentiate this function as it is currently
written. We would have to apply the quotient rule once and the product rule twice, among
other things. We could do some algebra and multiply out some of the factors in the numerator
and the denominator of f (x) to make our job easier, but that is in itself a pretty nasty calculation.
However, it happens to be not nearly as difficult to differentiate the related function ln | f (x)|.
Taking the logarithm of both sides and then differentiating both sides, we have
√
x (x 2 − 1)5
(x + 2)(x − 4)3
√ 2
x(x − 1)5 ln |y| = ln (x + 2)(x − 4)3 √
ln |y| = ln | x| + ln |(x 2 − 1)5 | − ln |x + 2| − ln |(x − 4)3 |
y=
ln |y| =
1
ln |x| + 5 ln |x 2 − 1| − ln |x + 2| − 3 ln |x − 4|
2
d
d
( ln |y|) =
dx
dx
1
ln |x| + 5 ln |x 2 − 1| − ln |x + 2| − 3 ln |x − 4|
2
1
1 5(2x)
1
3
y =
+ 2
−
−
y
2x x − 1 x + 2 x − 4
1
10x
1
3
y =y
+
−
−
2x x 2 − 1 x + 2 x − 4
√ 2
x(x − 1)5
10x
1
3
1
y =
+
−
−
.
(x + 2)(x − 4)3 2x x 2 − 1 x + 2 x − 4
← set y = f (x)
← apply ln |x|
← algebra
← algebra
← differentiate
← derivative rules
← solve for y
← definition of y
CONFIRMING PAGES
13:1
TKmaster2010
WHF00153/FREE087-Taalman
220
Chapter 2
November 21, 2012
Derivatives
Although the calculation as a whole is algebraically complicated, the differentiation step
was simple. The most difficult part was to remember to use the chain rule when we differd
1
entiated ( ln y) = y . Note also that we used the fact that |ab| = |a||b| and the fact that
dx
y
|ab | = |a|b in our calculations.
EXAMPLE 5
Using logarithms to differentiate a function with variables in both base
and exponent
Use logarithmic differentiation to find
d
((x 2
dx
− 3) x ).
SOLUTION
In this example using logarithmic differentiation is not a choice, but a necessity: Neither
the product rule nor the exponential rule applies to this function. When taking the derivative of a function that involves the variable in both the base and the exponent, we must
use logarithmic differentiation. After setting y = (x 2 − 3) x and then applying the natural
logarithm to both sides, we will be able to use algebra to remove the variable x from the
exponent and then differentiate both sides:
y = (x 2 − 3) x
← set y = f (x)
ln y = ln((x 2 − 3) x )
← apply ln(x) to both sides
ln y = x ln(x 2 − 3)
← algebra
d
d
( ln y) = (x ln(x 2 − 3))
dx
dx
← differentiate both sides
2x
1 y = (1) ln(x 2 − 3) + (x) 2
y
x −3
2x 2
y = y ln(x 2 − 3) + 2
x −3
2x 2
y = (x 2 − 3) x ln(x 2 − 3) + 2
.
x −3
← chain, product rules
← solve for y
← since y = (x 2 − 3) x
Note that we did not need any absolute values in this calculation, because (x 2 − 3) x is
always positive where it is defined.
?
TEST YOUR
UNDERSTANDING
What differentiation fact is the consequence of the limit statement lim
h→0
eh − 1
h
= 1 that
characterizes the number e?
In the proof that
d
(ln x)
dx
1
x
= , we used the fact that e ln x = x. It is also true that ln(e x ) =
x; could we have started with this equality instead? Why or why not?
What can you say about a quantity that grows at a rate proportional to the amount of
the quantity that is present?
Although there are absolute values in the fifth line of the calculation in Example 4,
there are no absolute values in the sixth line; what happened to the absolute values and
why?
Why was logarithmic differentiation necessary in Example 5? In particular, why did
neither the power nor the exponential rule apply?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
2.5
221
Derivatives of Exponential and Logarithmic Functions
EXERCISES 2.5
Thinking Back
Solving exponential and logarithmic equations: Use rules of exponents and logarithms to solve each of the following equations.
3(1.2) x = 500
ln(x + 1)
=0
ln(x − 2)
ln(x 2 + x − 5) = 0
2x + 1
=0
3x − 5
Compositions: For each function k, find functions f , g, and h
such that k = f ◦ g ◦ h. There may be more than one possible
answer.
√
√
k(x) = e 1/ x+1
k(x) = ln( x 2 + 5 )
k(x) = 3 ln(5 x + 2)
k(x) = ( ln 3x)2
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
d π
(e ) = 0.
dx
d z
(b) True or False: (e ) = e z .
dz
d 1
= ln x.
(c) True or False:
dx x
d
1
(d) True or False: (ln |x|) = .
dx
|x|
(a) True or False:
(e) True or False: If f is an exponential function, then f is
a constant multiple of f .
(f) True or False: If f is a constant multiple of f , then f is
an exponential function.
6. The function f (x) = e x is its own derivative. Are there
other functions with this property? If not, explain why
not. If so, give three examples.
7. Explain how the formula for differentiating the natural
logarithm function is a special case of the formula for differentiating logarithmic functions of the form logb x.
d
1
( ln x) = , we really mean to condx
x
1
sider the function
on the restricted domain (0, ∞).
x
8. When we say that
Why?
9. The graphs of the exponential functions y = 2 x , y = 4 x ,
and y = 2(2 x ) are shown in the figure at the left. Use your
knowledge of transformations to determine which graph
is which without using a graphing calculator.
y
(g) True or False: Logarithmic differentiation is required
in order to differentiate complicated products and
quotients.
(h) True or False: Logarithmic differentiation is required
in order to differentiate expressions that have a variable in both the base and the exponent.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Three functions f whose derivatives are just constant
multiples of f .
(b) Three functions that are transcendental, but whose
derivatives are algebraic.
(c) A function whose derivative would be difficult or impossible to find without the method of logarithmic
differentiation.
3. Does the exponential rule apply to the function f (x) = x x ?
What about the power rule? Explain your answers.
4. The natural exponential function is its own derivative.
Explain what this means graphically. (Use words like
“height” and “slope.”)
5. Explain how the formula for differentiating the natural exponential function is a special case of the formula for differentiating exponential functions of the form e kx . Then
explain why it is a special case of the formula for differentiating functions of the form b x .
y = e x , y = e 3x , y = e −2x
y = 2 x , y = 4 x , y = 2(2 x )
y
6
6
5
5
4
4
3
3
2
2
1
⫺2
⫺1
1
1
x
2
⫺2
⫺1
1
2
x
10. The graphs of the exponential functions y = e x , y = e 3x ,
and y = e −2x are shown in the preceding graph at
the right. Use your knowledge of transformations to
determine which graph is which without using a graphing
calculator.
2
11. The functions f (x) = 2(x ) and g(x) = (2 x )2 look similar,
but are very different functions. (This is why we try to
2
avoid using the ambiguous notation 2 x .) Calculate f (3)
and g(3), and show that they are not the same (and thus
that f (x) and g(x) are not the same function). Then find all
the values for which f (x) = g(x).
12. Every exponential function of the form f (x) = b x (with
b > 0 and b = 1) is one-to-one. Explain why this fact implies that b x = b y if and only if x = y.
13. Explain how we know that logarithmic functions are oneto-one. Why does this mean that A = B if and only if
logb A = logb B (assuming that A and B are positive)?
CONFIRMING PAGES
13:7
TKmaster2010
WHF00153/FREE087-Taalman
222
Chapter 2
November 21, 2012
Derivatives
14. What is the definition of the number e? What does
eh − 1
? Why is this
this definition tell you about lim
h
h→0
limit relevant to calculating the derivative of the function
f (x) = e x ?
16. Why do we need to consider absolute values when
we apply logarithmic differentiation to f (x) = xe x sin x?
In contrast, why do we not need to consider absolute values when we apply logarithmic differentiation to
f (x) = x x ?
15. Describe the process called logarithmic differentiation.
What types of differentiation problems is logarithmic differentiation useful for?
Skills
Find the derivatives of each of the functions in Exercises
17–44. In some cases it may be convenient to do some preliminary algebra.
17. f (x) =
1
2 − e 5x
19. f (x) = 3x 2 e −4x
21. f (x) =
1−x
ex
18. f (x) = log2 (3x 2 − 5)
47. f (x) =
20. f (x) = e 3x ln(x 2 + 1)
x3
22. f (x) = ln 2
x +x+1
48. f (x) =
e x ln x
x2 − 1
23. f (x) = e x (x 2 +3x−1)
24. f (x) =
25. f (x) = e
26. f (x) = 3 + log3 x
x
3 ln x
27. f (x) = e (e
x
28. f (x) = e (x
)
29. f (x) = (e )
31. f (x) =
e
)
30. f (x) = (x )
x e
e e
5
ln(x )
ln(x 4 )
32. f (x) = 11 + e π − ln 2
−3 2x
37. f (x) = ln(x + e )
39. f (x) = log2 (3 x − 5)
34. f (x) = ln(x 2 )
36. f (x) = ln(x 2 + 1)
√
ex
38. f (x) = √
ln x
40. f (x) = ln(x 2 +1)(e x )− 1/3
41. f (x) = x 2 ln( ln x)
42. f (x) = ln(x x )
43. f (x) = (2 x ) x
44. f (x) = 21−3
33. f (x) = x
2 x
e
35. f (x) = x 2 log2 (x2 x )
2
√
x
46. f (x) =
x
Describe the derivatives of each of the piecewise-defined
functions in Exercises 45–48.
⎧ x
⎨ 2 , if x ≤ −2
45. f (x) = 1
⎩ , if x > −2
x2
ln(−x), if x < 0
ln x, if x ≥ 0
x 2 , if x < 1
1 + ln x, if x ≥ 1
⎧
1
1
⎪
⎪
⎨ − x, if x ≤ 0
2
⎪
⎪
⎩
4
1
, if x > 0
1 + ex
Use logarithmic differentiation to find the derivatives of each
of the functions in Exercises 49–58.
√
√
50. f (x) = 12x 3 1 − x x + 1
49. f (x) = x ln |2 x + 1|
√
e 2x (x 3 − 2)4
2x x3 − 1
52. f (x) =
51. f (x) = √
x(3e 5x + 1)
x(2x − 1)
53. f (x) = xln x
x x
55. f (x) =
x−1
57. f (x) = ( ln x)ln x
54. f (x) = (2x + 1)3x
56. f (x) = ( ln x) x
x
1
58. f (x) =
x+1
In Exercises 59–63, find a function f that has the given derivative f . In each case you can find the answer with an educated
guess-and-check process.
59. f (x) =
4e 4x (3x 5 − 1) − e 4x (15x 4 )
(3x 5 − 1)2
60. f (x) = x 2 e x
3
62. f (x) = e x (1 + e x )
x
x2 + 3
ex
63. f (x) =
1 + ex
61. f (x) =
Applications
64. An abandoned building contained 45 rats on the first day
of the year and 53 rats 30 days later. Let r(t) be the function that describes the number of rats in the building
t days after the first of the year.
(a) Find a formula for r(t) given that the rate of change
of the rat population is constant, and use this formula
to predict the number of rats in the building on the
100th day of the year (t = 99).
(b) Find a formula for r(t) given that the rate of change
of the rat population is proportional to the number of
rats in the building, and use this formula to predict
the number of rats in the building on the 100th day
of the year (t = 99).
65. Alina started an investment account with an initial
deposit of one thousand dollars. Following the initial
deposit, the amount of money increased at a rate proportional to her investment account balance. After three
years her balance was $1,260.
(a) Write down a function A(t) that describes the amount
of money in Alina’s investment account t years after
her initial deposit.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 10, 2012
2.5
(b) How much money will Alina have in her investment
account after 30 years?
(c) How long will it be before Alina’s initial investment
quadruples?
66. The temperature T, in degrees Fahrenheit, of a yam after
sitting in a hot oven for t minutes is given by the function
Derivatives of Exponential and Logarithmic Functions
related to the odd saying “Cold water boils faster.”
How?
67. A political candidate starts an advertising campaign in
Hamtown, Virginia. The number of people in Hamtown
that have heard of him t days after the start of his campaign is given by
T(t) = 350 − 280e −0.2t .
P(t) =
(a) What is the initial temperature of the yam, before it
is put in the oven?
(b) Given that over time the temperature of the yam will
approach the temperature inside the oven, use a limit
to determine the temperature of the oven.
(c) How long will it take for the yam to be within
5 degrees Fahrenheit of the temperature of the oven?
(d) The first derivative of T(t) measures the rate of
change of the temperature of the yam. The second
derivative of T(t) measures the rate of change of the
rate of change of the temperature of the yam. Use
T (t) and T (t) to argue that the temperature of the
yam increases at a decreasing rate. This statement is
223
45,000
.
1 + 35e −0.12t
(a) How many people knew about the candidate before
the start of his advertising campaign?
(b) Given that over time the advertising will eventually
reach everyone in the town, use a limit to determine
the population of Hamtown.
(c) How many days will it take for all but one person in
Hamtown to have heard of the candidate?
(d) Find P (t) and use it to argue that in this model, the
number of people who have heard of the candidate is
always increasing. Does this make sense in the context of this problem?
Proofs
68. Use the definition of the derivative and the definition of
the number e to prove that f (x) = e x is its own derivative.
69. Use the chain rule to prove that
d kx
(e ) = ke kx .
dx
d
d
70. Use the fact that (b x ) = ( ln b) b x to prove that (e kx ) =
dx
dx
ke kx .
71. Prove that if f is an exponential function, then f (x) is
proportional to f (x).
72. Use implicit differentiation and the fact that logb x is the
inverse of b x to prove that
73. Use the definition of |x|, the chain rule, and the fact that
d
1
d
1
( ln x) = for x > 0 to prove that ( ln |x|) = for all
dx
x
dx
x
x = 0.
74. Use a direct application of the fact that
1
f ( f −1 (x))
to prove that
d
( f −1 (x)) =
dx
d
1
( ln x) = .
dx
x
d
1
(logb x) =
.
dx
(lnb)x
Thinking Forward
L’Hôpital’s rule: At the end of Chapter 3 we will see that, under certain conditions, the limit of a quotient of functions is
equal to the limit of the quotient of the derivatives of those
functions. Specifically, if f (x) and g(x) both approach zero as
x → c, then
lim
x→c
f (x)
f (x)
= lim .
g(x) x→c g (x)
lim
x3
1 − 2x
lim
3 −3
1 − x2
lim
ln x
x−1
lim
(x − 3)2
1 − e x−3
x→1
x→3
= ky. A solution of a differential equation is a
Show that y(x) = 4e 3x is a solution of the differential
equation
then use L’Hôpital’s rule to calculate the limit:
x→1
dx
function y(x) that makes the equation true.
0
Show that each of the following limits is of the form and
0
x→0
dy
of the form
dx
= 3y.
Show that y(x) = 1.7e −2.1x is a solution of the differential equation
x
Differential equations: A function y(x) is exponential if and only
if its derivative is proportional to itself. This means that exponential functions are solutions of differential equations
dy
= −2.1y.
dy
dx
= ( ln 2)y.
Describe all of the solutions of the differential equation
dx
Show that y(x) = 3(2 x ) is a solution of the differential
equation
dy
dy
dx
= 3y.
Describe all of the solutions of the differential equation
dy
dx
= 3y that satisfy y(0) = 2.
CONFIRMING PAGES
13:26
TKmaster2010
WHF00153/FREE087-Taalman
224
2.6
Chapter 2
November 21, 2012
Derivatives
DERIVATIVES OF TRIGONOMETRIC AND HYPERBOLIC*
FUNCTIONS
Derivatives of the six trigonometric functions
Derivatives of inverse trigonometric functions
Hyperbolic functions, inverse hyperbolic functions, and their derivatives
Derivatives of Trigonometric Functions
Because trigonometric functions have periodic oscillating behavior, and their slopes also
have periodic oscillating behavior, it would make sense if the derivatives of trigonometric
functions were trigonometric. For example, the two graphs that follow show the function
f (x) = sin x and its derivative f (x) = cos x. As we will prove in Theorem 2.17, it turns out
that, at each value of x, the slope of the graph of f (x) = sin x is given by the height of the
graph of f (x) = cos x. Before we tackle this fact algebraically, take a minute to verify that
π
it is the case with these graphs for the values x = −5.2, x = , and x = 4, as shown in the
2
following figures:
Slopes of f (x) = sin x at three points
Heights of f (x) = cos x at three points
y
y
1
1
π
2
5.2
4
x
π
2
5.2
4
x
1
1
The six trigonometric functions have the following derivatives:
THEOREM 2.17
Derivatives of the Trigonometric Functions
For all values of x at which the following functions below are defined,
(a)
d
(sin x) = cos x
dx
(c)
d
(tan x) = sec2 x
dx
(e)
d
(cot x) = − csc2 x
dx
(b)
d
(cos x) = − sin x
dx
(d)
d
(sec x) = sec x tan x
dx
(f)
d
(csc x) = − csc x cot x
dx
It is important to note that these formulas for derivatives are true only if angles are measured in radians; see Exercise 5.
Proof. We will prove the formulas for sin x and tan x from parts (a) and (c) and leave the proofs of
the remaining four formulas to Exercises 83–86.
(a) The proof of the first formula is nothing more than an annotated calculation using the definition of the derivative. To simplify the limit we obtain, we will rewrite sin(x + h) with a
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
225
Derivatives of Trigonometric and Hyperbolic Functions
trigonometric identity. Our goal after that will be to rewrite the limit so that we can apply the
two trigonometric limits from Theorem 1.35 in Section 1.6. The calculation is as follows:
sin(x + h) − sin x
d
(sin x) = lim
dx
h→0
h
= lim
h→0
← definition of the derivative
(sin x cos h + sin h cos x) − sin x
h
← sum identity for sine
sin x(cos h − 1) + sin h cos x
h
sin h
cos h − 1
= lim sin x
+ cos x
h→0
h
h
cos h − 1
sin h
+ cos x lim
= sin x lim
h→0
h→0 h
h
= lim
h→0
= (sin x)(0) + (cos x)(1) = cos x.
← algebra
← algebra
← limit rules
← trigonometric limits
(c) We do not have to resort to the definition of the derivative in order to prove the formula for
differentiating tan x. Instead we can use the quotient rule, the fact that tan x =
sinx
, and the
cosx
formulas for differentiating sin x and cos x:
d
d sin x
(tan x) =
dx
dx cos x
d
d
(sin x) · (cos x) − (sin x) · (cos x)
dx
dx
=
2
(cos x)
← quotient rule
=
(cos x)(cos x) − (sin x)(− sin x)
cos2 x
← derivatives of sin x and cos x
=
1
cos2 x + sin2 x
=
= sec2 x.
cos2 x
cos2 x
← algebra and identities
Derivatives of Inverse Trigonometric Functions
We can use the formulas for the derivatives of the trigonometric functions to prove the
following formulas for the derivatives of the inverse trigonometric functions:
THEOREM 2.18
Derivatives of Inverse Trigonometric Functions
For all values of x at which the following functions are defined,
(a)
1
d
(sin−1 x) = √
dx
1 − x2
(b)
1
d
(tan−1 x) =
dx
1 + x2
(c)
1
d
(sec−1 x) = √
dx
|x| x 2 − 1
It is extremely important and surprising to note that although inverse trigonometric functions are transcendental, their derivatives are algebraic! This property makes these derivative formulas particularly useful for finding certain antiderivatives, and in Chapter 6 they
will be part of our arsenal of integration techniques. Of course, all of these rules can be
used in combination with the sum, product, quotient, and chain rules. For example,
1
d
(sin−1 (3x 2 − 1)) = (6x).
dx
1 − (3x 2 − 1)2
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
226
Chapter 2
November 21, 2012
Derivatives
Proof. We will prove the rule for sin−1 x and leave the remaining two rules to Exercises 87 and 88.
We could apply Theorem 2.16 here, but it is just as easy to do the implicit differentiation by hand.
Since sin(sin−1 x) = x for all x in the domain of sin−1 x, we have
sin(sin−1 x) = x
← sin−1 x is the inverse of sin x
d
d
(sin(sin−1 x)) = (x)
dx
dx
d
cos(sin−1 x) · (sin−1 x) = 1
dx
← differentiate both sides
← chain rule
1
d
(sin−1 x) =
dx
cos(sin−1 x)
1
d
(sin−1 x) = dx
1 − sin2 (sin−1 x)
1
d
(sin−1 x) = √
.
dx
1 − x2
← algebra
← since sin2 x + cos2 x = 1
← sin x is the inverse of sin−1 x
We could also have used triangles√and the unit circle to show that the composition cos(sin−1 x) is
equal to the algebraic expression 1 − x 2 , as we did in Example 4 of Section 0.4.
An interesting fact about the derivatives of the inverse sine and inverse secant functions
is that their domains are slightly smaller than the domains of the original functions. The
graphs of the inverse trigonometric functions are as follows (note their domains):
f (x) = sin−1 x
has domain [−1, 1]
g(x) = tan−1 x
has domain (−∞, ∞)
y
h(x) = sec−1 x
has domain (−∞, −1] ∪ [1, ∞)
y
y
π
π
π
π
2
π
2
π
2
1
1
x
1
1
π
2
π
π
x
1
1
2
x
2
π
If you look closely at the first and third graphs, you should notice that at the ends of
the domains the tangent lines will be vertical. Since a vertical line has undefined slope,
the derivative does not exist at these points. This means that the derivatives of sin−1 x and
sec−1 x are not defined at x = 1 or x = −1; see the first and third graphs shown next:
1
f (x) = √
1 − x2
has domain (−1, 1)
y
1
√
|x| x 2 − 1
has domain (−∞, −1) ∪ (1, ∞)
y
1
1 + x2
has domain (−∞, ∞)
y
h (x) =
g (x) =
3
3
3
2
2
2
1
1
1
1
1
1
x
1
1
x
1
1
1
CONFIRMING PAGES
1
x
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
Derivatives of Trigonometric and Hyperbolic Functions
227
Hyperbolic Functions and Their Derivatives*
The trigonometric functions sine and cosine are circular functions in the sense that they
are defined to be the coordinates of a parameterization of the unit circle. This means that
the circle defined by x 2 +y2 = 1 is the path traced out by the coordinates (x, y) = (cos t, sin t)
as t varies; see the following figure at the left:
Points on the circle x 2 + y 2 = 1
Points on the hyperbola x 2 − y 2 = 1
y
y
2
1
2
1
2
(x, y) (cos t, sin t)
1
2
1
x
2
1
1
1
1
2
2
(x, y) (cosh t, sinh t)
x
2
Now let’s consider the path traced out by the hyperbola x 2 − y2 = 1 as shown at the right.
One parameterization of the right half of this hyperbola is traced out by the hyperbolic
functions (cosh t, sinh t) that we will spend the rest of this section investigating.
The hyperbolic functions are nothing more than simple combinations of the exponential functions e x and e −x :
DEFINITION 2.19
Hypberbolic Sine and Hyperbolic Cosine
For any real number x, the hyperbolic sine function and the hyperbolic cosine function
are, respectively, defined as the following combinations of exponential functions:
sinh x =
e x − e −x
2
cosh x =
e x + e −x
2
The hyperbolic sine function is pronounced “sinch” and the hyperbolic cosine function
is pronounced “cosh.” The “h” is for “hyperbolic.” As we will soon see, the properties
and interrelationships among the hyperbolic functions are similar to the properties and
interrelationships among the trigonometric functions. These properties will be particularly
useful in Chapter 6 when we attempt to solve certain forms of integrals.
It is a simple matter to use Definition 2.19 to verify that, for any value of t, the point
(x, y) = (cosh t, sinh t) lies on the hyperbola x 2 − y2 = 1; see Exercise 89. We will usually
think of this fact rewritten so that the independent variable is x, as follows:
cosh2 x − sinh2 x = 1.
Here we are using the familiar convention that, for example, sinh2 x is shorthand for
(sinh x)2 . Note the similarity between the hyperbolic identity cosh2 t − sinh2 t = 1 and
the Pythagorean identity for sine and cosine. Hyperbolic functions also satisfy many other
algebraic identities that are reminiscent of those that hold for trigonometric functions, as
you will see in Exercises 90–92.
Just as we can define four additional trigonometric functions from sine and cosine, we
can define four additional hyperbolic functions from hyperbolic sine and hyperbolic cosine.
We will be interested primarily in the hyperbolic tangent function:
tanh x =
sinh x
e x − e −x
.
= x
cosh x
e + e −x
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
228
Chapter 2
November 21, 2012
Derivatives
We can also define csch x, sech x, and coth x as the reciprocals of sinh x, cosh x, and tanh x,
respectively.
The graphs of sinh x, cosh x, and tanh x are shown next. In Exercises 13–16 you will
investigate various properties of these graphs.
y = sinh x
y = cosh x
y
y
8
8
4
6
3 2 1
y = tanh x
1
2
3
y
1
4
x
3 2 1
1
2
3
x
2
4
8
3 2 1
1
2
3
1
x
In Chapter 4 we will see that the graph of y = cosh x is an example of a catenary curve
(see also Exercise 82), which is the shape formed by a hanging chain or cable.
As with any functions that we study, we are interested in finding formulas for the
derivatives of sinh x, cosh x, and tanh x. The similarity between hyperbolic functions and
trigonometric functions continues here. These derivatives follow a very familiar pattern,
differing from the pattern for trigonometric functions only by a sign change.
THEOREM 2.20
Derivatives of Hyperbolic Functions
For all real numbers x,
(a)
d
(sinh x) = cosh x
dx
(b)
d
(cosh x) = sinh x
dx
(c)
d
(tanh x) = sech2 x
dx
If you prefer to stay away from the hyperbolic secant function sech x, you can write the
1
derivative in part (c) as
.
2
cosh x
Proof. The proofs of these differentiation formulas follow immediately from the definitions of the
hyperbolic functions as simple combinations of exponential functions. For example,
d
d 1 x
1
(sinh x) =
(e − e −x ) = (e x + e −x ) = cosh x.
dx
dx
2
2
The proofs of parts (b) and (c) are left to Exercises 93 and 94.
Although hyperbolic functions may seem somewhat exotic, they work with the other
differentiation rules just as any other functions do. For example, with the product and chain
rules we can calculate
d
(5x sinh3 x 2 ) = 5 sinh3 x 2 + 5x(3 sinh2 x 2 )(cosh x 2 )(2x).
dx
The derivatives of the hyperbolic cotangent, secant, and cosecant functions are also similar
to those of their trigonometric cousins, but at the moment we will be focusing only on
hyperbolic sine, cosine, and tangent.
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
Derivatives of Trigonometric and Hyperbolic Functions
229
Inverse Hyperbolic Functions and Their Derivatives*
For a function to have an inverse, it must be one-to-one. Looking back at the graphs of
sinh x, cosh x, and tanh x, we see that only cosh x fails to be one-to-one. Just as we did
when we defined the trigonometric inverses, we will restrict the domain of cosh x to a
smaller domain on which it is one-to-one. We will choose the restricted domain of cosh x
to be x ≥ 0. The notation we will use for the inverses of these three functions is what you
would expect: sinh−1 x, cosh−1 x and tanh−1 x.
Since the hyperbolic functions are defined as combinations of exponential functions, it
would seem reasonable to expect that their inverses could be expressed in terms of logarithmic functions. This is in fact the case, as you will see in Exercises 97–99. However, our
main concern here is to find formulas for the derivatives of the inverse hyperbolic functions,
which we can do directly from identities and properties of inverses.
THEOREM 2.21
Derivatives of Inverse Hyperbolic Functions
For all x for which the following are defined,
(a)
1
d
(sinh−1 x) = √
2
dx
x +1
(b)
1
d
(cosh−1 x) = √
2
dx
x −1
(c)
1
d
(tanh−1 x) =
dx
1 − x2
Similar formulas can be developed for the inverse hyperbolic cotangent, secant, and cosecant functions. Notice the strong similarities between these derivatives and the derivatives
of the inverse trigonometric functions.
Proof. We will prove the rule for the derivative of sinh−1 x and leave the remaining two rules to
Exercises 95 and 96. Starting from the fact that sinh(sinh−1 x) = x for all x, we can apply implicit
differentiation:
sinh(sinh−1 x) = x
← sinh−1 x is the inverse of sinh x
d
d
(sinh(sinh−1 x)) = (x)
dx
dx
cosh(sinh−1 x) ·
← differentiate both sides
d
(sinh−1 x) = 1
dx
← chain rule, derivative of sinh x
1
d
(sinh−1 x) =
.
dx
cosh(sinh−1 x)
1
d
(sinh−1 x) = dx
2
← algebra
1
d
(sinh−1 x) = √
.
dx
1 + x2
← sinh x is the inverse of sinh−1 x
1 + sinh (sinh−1 x)
← since cosh2 x − sinh2 x = 1
Compare this proof with our proof earlier in this section for the derivative of sin−1 x; the two are
similar.
Examples and Explorations
EXAMPLE 1
Differentiating combinations of trigonometric functions
Find the derivatives of each of the following functions:
(a) f (x) =
tan x
x3 − 2
(b) f (x) = x sin−1 (3x + 1)
(c) f (x) = sec2 e x
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
230
Chapter 2
December 10, 2012
Derivatives
SOLUTION
tanx
(a) The function f (x) = 3
is a quotient of two functions. By the quotient rule and the
x −2
rule for differentiating tangent, we have
d
dx
tan x
x3 − 2
=
=
d
d
(tan x) · (x 3 − 2) − (tan x) · (x 3 − 2)
dx
dx
(x 3 − 2)2
(sec2 x)(x 3
− 2) − (tan x)(3x 2 )
.
(x 3 − 2)2
(b) The function f (x) = x sin−1 (3x + 1) is a product of two functions, and thus we begin
with the product rule. We will also need the chain rule to differentiate the composition
sin−1 (3x + 1):
1
f (x) = (1) · sin−1 (3x + 1) + x · (3)
1 − (3x + 1)2
3x
.
= sin−1 (3x + 1) + 1 − (3x + 1)2
(c) The function f (x) = sec2 e x is a composition of three functions, and thus we need to
apply the chain rule twice:
d
d
(sec2 e x ) = ((sec(e x ))2 )
dx
dx
= 2(sec e x )1 ·
← rewrite so compositions are clear
d
(sec e x )
dx
= 2(sec e x )(sec e x )(tan e x ) ·
= 2(sec e x )(sec e x )(tan e x )e x
← first application of chain rule
d x
(e )
dx
← second application of chain rule
← derivative of e x
Perhaps the most difficult part of the preceding calculation is that the derivative of sec x
d
has two instances of the independent variable: (sec x) = sec x tan x. This means that
dx
we needed to put the “inside” function e x into both of the slots for variables.
EXAMPLE 2
Differentiating combinations of hyperbolic functions*
Find the derivatives of each of the following functions:
(a) f (x) = ln(tanh2 (x 3 + 2x + 1)) (b) f (x) = cosh−1 (e 3x )
SOLUTION
(a) This is a nested chain-rule problem, since f (x) is a composition of multiple functions.
We will work from the outside to the inside, one step at a time:
f (x) =
1
d
(tanh2 (x 3 + 2x + 1))
+ 2x + 1) dx
1
d
(2 tanh(x 3 + 2x + 1)) (tanh(x 3 + 2x + 1))
=
2 3
dx
tanh (x + 2x + 1)
1
(2 tanh(x 3 + 2x + 1))(sech2 (x 3 + 2x + 1))(3x 2 + 2).
=
2 3
tanh (x + 2x + 1)
2
tanh
(x 3
(b) Once again we have a nested chain-rule situation. Notice in particular how e 3x works
with the derivative of the inverse hyperbolic cosine function:
1
1
1
−1 3x − 1/2 d
−1 3x
−1 3x − 1/2
f (x) = (cosh (e ))
(3e 3x ).
(cosh (e )) = (cosh (e ))
2
dx
2
(e 3x )2 − 1
CONFIRMING PAGES
13:36
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
EXAMPLE 3
Derivatives of Trigonometric and Hyperbolic Functions
231
Finding antiderivatives that involve inverse trigonometric functions
1
Find a function f whose derivative is f (x) =
.
1 + 4x 2
SOLUTION
Since the derivative of tan−1 x is
1
,
1 + x2
we might suspect that the function we are looking
for is related to the inverse tangent function. We will use an intelligent guess-and-check
method to find f . Clearly f (x) = tan−1 x isn’t exactly right, since its derivative is missing the
“4.” A good guess might be f (x) = tan−1 (4x); let’s try that:
1
4
d
(tan−1 (4x)) =
(4) =
.
dx
1 + (4 x)2
1 + 16x 2
Obviously that wasn’t quite right either; but by examining the results we can make a new
guess. We might try tan−1 (2x), since the “2x” will be squared in the derivative and become
the “4x 2 ” we are looking for in the denominator:
1
2
d
(tan−1 (2x)) =
(2) =
.
2
dx
1 + (2x)
1 + 4x2
Now we are getting somewhere; this result differs by a multiplicative constant from the
derivative f (x) we are looking for, and that is easy to fix. We need only divide our guess by
1
that constant. Try the function f (x) = tan−1 (2x):
2
1
1
d 1
1
tan−1 (2x) =
(2) =
.
dx 2
2 1 + (2x)2
1 + 4x 2
We now know that f (x) =
1
2
tan−1 (2x) is a function whose derivative is f (x) =
1
.
1 + 4x 2
Of
course, we could also add any constant to f (x) and not change its derivative; for example, f (x) =
1
2
EXAMPLE 4
1
2
tan−1 (2x) + 5 would work as well. In fact, any function of the form f (x) =
tan−1 (2x) + C will have f (x) =
1
.
1 + 4x 2
Finding antiderivatives that involve hyperbolic functions*
ex
Find a function f whose derivative is f (x) = √
.
e 2x − 1
SOLUTION
Until we learn more specific antidifferentiation techniques in Chapter 6, a problem like this
is best done by an intelligent guess-and-check procedure. Given that we have the inverse
hyperbolic functions in mind, the best match of the three is the derivative of cosh−1 x. Since
the expression for f (x) also involves an e x , let’s revise that guess right away to cosh−1 e x .
Now we check by differentiating with the chain rule:
x
1
e
d
· ex = √
(cosh−1 e x ) = .
2x
dx
x
2
e −1
(e ) − 1
We guessed it on the first try! We have just shown that f (x) = cosh−1 e x has the desired
derivative.
TEST YOUR
? UNDERSTANDING
What trigonometric limits were used to find the derivative of sin x?
How can we obtain the derivative of sec x from the derivative of cos x?
What is the graphical reason that the domains of the derivatives of sin−1 x and sec−1 x
are slightly smaller than the domains of the functions themselves?
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
232
Chapter 2
December 10, 2012
Derivatives
How are hyperbolic functions similar to trigonometric functions? How are they
different?
How can we obtain the derivative of sinh−1 x from the derivative of sinh x?
EXERCISES 2.6
Thinking Back
Trigonometric and inverse trigonometric values: Without using
a calculator, find the exact values of each of the following
quantities.
π
π
sin −
tan −
sec
3
5π
6
√
tan−1 ( 3 )
4
sin−1 1
sec−1 (−2)
Compositions: For each function k that follows, find functions
f , g, and h such that k = f ◦ g ◦ h. There may be more than one
possible answer.
k(x) =
1
sin(x 3 )
k(x) = sin−1 (cos2 x)
k(x) = tan2 (3x + 1)
k(x) = sec(x 3 ) tan(x 3 )
Writing trigonometric compositions algebraically: Prove each
of the following equalities, which rewrite compositions of
trigonometric and inverse trigonometric functions as algebraic
functions.
√
√
cos(sin−1 x) = 1 − x 2 sin(cos−1 x) = 1 − x 2
sec2 (tan−1 x) = 1 + x 2
tan(sec−1 x) = |x| 1 −
1
x2
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: To find the derivative of sin x we have to
use the definition of the derivative.
(b) True or False: To find the derivative of tan x we have to
use the definition of the derivative.
(c) True or False: The derivative of
x4
4x 3
is
.
sinx
cosx
(d) True or False: If a function is algebraic, then so is its
derivative.
(e) True or False: If a function is transcendental, then so
is its derivative.
(f) True or False: If f is a trigonometric function, then f is also a trigonometric function.
(g) True or False: If f is an inverse trigonometric function,
then f is also an inverse trigonometric function.
(h) True or False: If f is a hyperbolic function, then f is
also a hyperbolic function.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function that is its own fourth derivative.
(b) A function whose domain is larger than the domain
of its derivative.
(c) Three non-logarithmic functions that are transcendental, but whose derivatives are algebraic.
3. What limit facts and trigonometric identities are used in
the proof that
d
(sin x) = cos x?
dx
4. Sketch graphs of sin x and cos x on [−2π, 2π ].
(a) Use the graph of sin x to determine where sin x is
increasing and and where it is decreasing.
(b) Use the graph of cos x to determine where cos x is
positive and where it is negative.
(c) Explain why your answers to parts (a) and (b) suggest
that cos x is the derivative of sin x.
5. The differentiation formula
d
(sin x) = cos x is valid only
dx
if x is measured in radians. In this exercise you will
explore why this derivative relationship does not hold if
x is measured in degrees.
(a) Set your calculator to degree mode, and sketch a
graph of sin x that shows at least two periods. If the
derivative of sin x is cos x, then the slope of your graph
at x = 0 should be equal to cos 0 = 1. Use your graph
to explain why this is not the case when we use degrees. (Hint: Think about your graphing window scale.)
(b) Now set your calculator back to radians mode!
6. Suppose you wish to differentiate g(x) = sin2 (x)+cos2 (x).
What is the fastest way to do this, and why?
7. The derivatives of the function f (x) = cos(3x 2 ) that follow
are incorrect. What misconception occurs in each case?
(a) Incorrect: f (x) = (− sin x)(3x 2 ) + (cos x)(6x).
(b) Incorrect: f (x) = − sin(6x).
8. The derivatives of the function f (x) = cos(3x 2 ) that follow
are incorrect. What misconception occurs in each case?
(a) Incorrect: f (x) = − sin(3x 2 ).
(b) Incorrect: f (x) = − sin(3x 2 )(6x)(6).
9. In the proof that
d
1
(sin−1 x) = √
, we used the fact
dx
1 − x2
that sin(sin−1 x) = x. It is also true that sin−1 (sin x) = x;
could we have started with that inequality instead? Why
or why not?
10. Both of the following equations are true: tan(tan−1 x) = x
and tan−1 (tan x) = x. We can find the derivative of tan−1 x
by differentiating both sides of one of these equations
CONFIRMING PAGES
23:58
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
and solving for
d
(tan−1 x). Which one of the equations
dx
should we use, and why?
1
11. How can the derivative of sin−1 x be equal to both √
1 − x2
1
? Which expression is easier to use, and
cos(sin−1 x)
and
why?
12. The function sin−1 x is defined on [−1, 1], but its derivative √
1
1 − x2
233
Derivatives of Trigonometric and Hyperbolic Functions
is defined only on (−1, 1). Explain why the
tangent lines to the graph of y = sin−1 x do not exist at
x = ±1. (Hint: Think about the corresponding tangent lines
on the graph of the restricted sine function.)
13. The figure that follows at the left shows the graphs of
1
2
y = sinh x, y = cosh x, and y = e x . For each of the statements that follows, explain graphically why the statement
is true. Then justify the statement algebraically, using the
definitions of the hyperbolic functions.
statements that follows, explain graphically why the
statement is true. Then justify the statement algebraically,
using the definitions of the hyperbolic functions.
1
2
1
2
(a) cosh x = e x + e −x
cosh x
=1
(1/2)e −x
15. The figure that follows at the left shows the graphs of
(b)
lim
x→−∞
1 x
1
e , and y = − e −x . For each of
2
2
y = sinh x, y =
the statements that follows, explain graphically why the
statement is true. Then justify the statement algebraically,
using the definitions of the hyperbolic functions.
1
2
1
2
(a) sinh x = e x − e −x
(b)
lim
x→−∞
sinh x
=1
−(1/2)e −x
Graph for Exercise 15
Graph for Exercise 16
y
y
1
(a) sinh x ≤ e x ≤ cosh x for all x
2
3
sinh x
cosh x
(b) lim
= 1 and lim
=1
x→∞ (1/2)e x
x→∞ (1/2)e x
2
1
1
Graph for Exercise 13
Graph for Exercise 14
y
y
1
1
1
3
3
2
2
2
3
1
2
2
2
x
3 2 1
1
2
3
x
1
1
1
1
1
2
x
2
1
1
1
2
2
3
3
2
x
14. The preceding figure at the right shows the graphs
1
2
1
2
of y = cosh x, y = e x , and y = e −x . For each of the
16. The preceding figure at the right shows the graphs of
y = tanh x, y = 1, and y = −1. For each of the statements that follows, explain graphically why the statement
is true. Then justify the statement algebraically, using the
definitions of the hyperbolic functions.
(a) −1 ≤ tanh x ≤ 1
(b) lim tanh x = 1 and lim tanh x = −1
x→∞
x→−∞
Skills
Find the derivatives of each of the functions in Exercises
17–50. In some cases it may be convenient to do some preliminary algebra.
x2 + 1
18. f (x) = 2 cos(x 3 )
17. f (x) =
cos x
19. f (x) = cot x − csc x
20. f (x) = tan2 (3x + 1)
3x 2 ln x
tan x
35. f (x) = sin( ln x)
ln(3x 2 )
tan x
36. f (x) = ln(x sin x)
37. f (x) = sin−1 (3x 2 )
38. f (x) = 3(sin−1 x)2
39. f (x) = x 2 arctan x 2
40. f (x) = tan−1 ( ln x)
33. f (x) =
34. f (x) =
21. f (x) = 4 sin2 x+4 cos2 x
22. f (x) = sec2 x −1
41. f (x) = sec−1 x 2
42. f (x) = sin(sin−1 x)
23. f (x) = 3 sec x tan x
24. f (x) = 3 sec x + 17
43. f (x) = sin−1 (sec2 x)
44. f (x) = sin2 (sec−1 x)
25. f (x) = sin(cos(sec(x)))
26. f (x) = csc (e )
27. f (x) = e
2
x
28. f (x) = e csc x
csc2 x
x
−2
5x sin x
√
31. f (x) = x sin x cos x
29. f (x) =
x
x
30. f (x) =
2
log3 (3 x )
sin2 x + cos2 x
sin x csc x
32. f (x) =
cot x cos x
45. f (x) =
sin−1 x
tan−1 x
46. f (x) =
sin−1 x
sec−1 x
47. f (x) = ln(arcsec (sin2 x))
48. f (x) = x −2 e 4x sin−1 x
49. f (x) = sec(1+tan−1 x)
50. f (x) =
CONFIRMING PAGES
sin(arcsin x)
arctan x
22:0
TKmaster2010
WHF00153/FREE087-Taalman
234
Chapter 2
November 27, 2012
Derivatives
Find the derivatives of each of the functions in Exercises
51–62. In some cases it may be convenient to do some preliminary algebra. (These exercises involve hyperbolic functions and
their inverses.)
51. f (x) = x sinh x 3
52. f (x) = x sinh3 x
53. f (x) = cosh(ln(x 2 + 1))
54. f (x) = 3 tanh2 e x
√
tanh x
56. f (x) = √
sinh x
55. f (x) =
cosh2 x + 1
57. f (x) = sinh−1 (x 3 )
59. f (x) =
58. f (x) = tanh−1 (tan x 2 )
60. f (x) = x tanh−1 x
sinh−1 x
cosh−1 x
61. f (x) = sin(e sinh
−1
x
62. f (x) = cosh−1 (cosh−1 x)
)
Use logarithmic differentiation to find the derivatives of each
of the functions in Exercises 63–65.
63. (sin x) x
64. (sec x) x
65. (sin x)cos x
In Exercises 66–71, find a function f that has the given derivative f . In each case you can find the answer with an educated
guess-and-check process.
2x
2
67. f (x) = √
66. f (x) = √
2
1 − 4x
1 − 4x 2
3x
1
69. f (x) =
68. f (x) =
1 + 9x 2
1 + 9x 2
70. f (x) =
3
71. f (x) = √
4 − 9x 2
1
9 + x2
In Exercises 72–77, find a function f that has the given derivative f . In each case you can find the answer with an educated guess-and-check process. (Some of these exercises involve
hyperbolic functions.)
2x
2
73. f (x) = √
72. f (x) = √
1 + 4x 2
1 + 4x 2
3x
1
75. f (x) =
74. f (x) =
1 − 9x 2
1 − 9x 2
76. f (x) =
3
77. f (x) = √
4 + 9x 2
1
9 − x2
Applications
78. In Exercise 83 from Section 1.6 we saw that the oscillating
position of a mass hanging from the end of a spring, neglecting air resistance, is given by the following equation,
where A, B, k, and m are constants:
s(t) = A sin
k
t + B cos
m
k
t .
m
(a) Show that the function s(t) has the property that
s (t) +
f k
s (t) + s(t) = 0 for some constant f . This is
m
m
the differential equation for spring motion, taking air
resistance into account. (Hint: Find the first and second
derivatives of s(t) first, and then show that s(t), s (t), and
s (t) have the given relationship.)
(b) Suppose the spring is released from an initial position of s0 with an initial velocity of v0 . Show that
2mv0 + f s0
A= spring coefficient, k
s(t)
s0
v0
mass, m
4km − f 2
and B = s0 .
80. Suppose your friend Max drops a penny from the top
floor of the Empire State Building, 1250 feet from the
ground. After t seconds, the penny is a distance of s(t) =
−16t 2 + 1250 from the ground. You are standing about a
block away, 250 feet from the base of the building.
(a) Show that the function s(t) has the property that
s (t) +
k
s(t) = 0. This is the differential equation for
m
the spring motion, an equation involving derivatives
that describes the motion of the bob on the end of the
spring.
(b) Suppose the spring is released from an initial position of s0 and with an initial velocity of v0 . Show that
A = v0
m
and B = s0 .
k
79. In Exercise 84 from Section 1.6 we learned that the oscillating position of a mass hanging from the end of a spring,
taking air resistance into account, is given by the following equation, where A, B, k, f , and m are constants:
4km − f 2
4km − f 2
t +B cos
t .
s(t) = e −f /2m t A sin
2m
1250 ft
␣(t)
250 ft
(a) Find a formula for the angle of elevation α(t) from the
ground at your feet to the height of the penny t seconds after Max drops it. Multiply by an appropriate
constant so that α(t) is measured in degrees.
2m
CONFIRMING PAGES
22:28
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
2.6
235
Derivatives of Trigonometric and Hyperbolic Functions
(b) Find a formula for the rate at which the angle of elevation α(t) is changing at time t, and use the formula
to determine the rate of change of the angle of elevation at the time the penny hits the ground.
81. The Gateway Arch in St. Louis is designed as an inverted
catenary curve. The arch is a complex three-dimensional
structure, but some sources model it simply, using the
hyperbolic function
1
(x − 299.22) ,
A(x) = 693.8 − 68.8 cosh
99.7
where x denotes the distance in feet from one base of the
arch as you approach the other. (This exercise involves
hyperbolic functions.)
y
600
(a) How tall is the arch, according to this model?
(b) There is a tram that takes visitors to an observation
deck in the top of the arch. The cabin of the tram rotates rather like cars on a Ferris wheel to keep visitors
upright, but the outer part of the tram changes angle
with the curve of the arch. What angle does the tram
make with the ground at the bottom of the arch?
(c) Visitors leave the tram at the observation deck 33 feet
from the center of the arch. What angle does the tram
make with the ground there?
82. Ian has climbed a pinnacle that is detached from the main
peak by roping down into the notch dividing them and
then climbing the pinnacle. He pulled an extra rope behind him so that he could get back to the main peak by
using a Tyrolean traverse, meaning that he would use the
rope to go directly back to the peak instead of descending and then climbing on rock again. When he anchors
the rope, it hangs in a catenary curve, with equation
r(x) = 125 cosh(0.008x − 0.6528).
400
The point x = 0 is where the rope attaches to the main
peak, while x = 136 is where it attaches to the pinnacle.
Heights are measured in feet above the notch. (This
exercise involves hyperbolic functions.)
200
200
400
x
600
y
y
y
y
x
x
x
33 ft
136
x
(a) How much higher is the main peak than the detached
pinnacle?
(b) Where is the low point of the rope as it hangs loosely?
How high is the rope above the notch at that point?
(c) What angle does the rope make with the horizontal
where it attaches to the main peak?
tram angle
tram angle
Proofs
83. Use the definition of the derivative, a trigonometric
identity, and known trigonometric limits to prove that
d
(cos x) = − sin x.
dx
that
84. Use the quotient rule and the derivative of the cosine
function to prove that
d
(sec x) = sec x tan x.
dx
85. Use the quotient rule and the derivative of the sine function to prove that
87. Use implicit differentiation and the fact that
tan(tan−1 x) = x for all x in the domain of tan−1 x to prove
d
(csc x) = − csc x cot x.
dx
d
1
(tan−1 x) =
.
dx
1 + x2
88. Use implicit differentiation and the fact that
sec(sec−1 x) = x for all x in the domain of sec−1 x to prove
that
d
1
(sec−1 x) = √ 2
. You will have to consider
dx
|x| x − 1
the cases x > 1 and x < −1 separately.
86. Use the quotient rule and the derivatives of the sine and
cosine functions to prove that
d
(cot x) = − csc2 x.
dx
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
236
Chapter 2
November 21, 2012
Derivatives
89. Prove that for any value of t, the point (x, y) =
(cosh t, sinh t) lies on the hyperbola x 2 − y2 = 1. Bonus
question: In fact, these points will always lie on the
right-hand side of the hyperbola; why? (This exercise involves hyperbolic functions.)
Use the definitions of the hyperbolic functions to prove that
each of the identities in Exercises 90–92 hold for all values of
x and y. Note how similar these identities are to those which
hold for trigonometric functions. (These exercises involve hyperbolic functions.)
90. (a) sinh(−x) = − sinh x, and (b) cosh(−x) = cosh x
91. sinh(x + y) = sinh x cosh y + cosh x sinh y
92. cosh(x + y) = cosh x cosh y + sinh x sinh y
Prove each of the differentiation formulas in Exercises 93–96.
(These exercises involve hyperbolic functions.)
93.
d
(cosh x) = sinh x
dx
94.
d
(tanh x) = sech2 x
dx
95.
1
d
(cosh−1 x) = √
2
dx
x −1
96.
1
d
(tanh−1 x)
dx
1 − x2
Prove that the inverse hyperbolic functions can be written in
terms of logarithms as shown in Exercises 97–99. (Hint for the
first problem: Solve sinh y = x for y by using algebra to get an
expression that is quadratic in e y (i.e., of the form ae 2y + be y + c)
and then applying the quadratic formula.)
(These exercises involve hyperbolic functions.)
√
97. sinh−1 x = ln(x + x 2 + 1 ), for any x.
√
98. cosh−1 x = ln(x + x 2 − 1 ), for x ≥ 1.
1+x
1
, for −1 < x < 1.
99. tanh−1 x = ln
2
1−x
Thinking Forward
Local extrema and inflection points: In the exercises that follow,
you will investigate how derivatives can help us find the locations of the maxima and minima of a function.
Suppose f has a maximum or minimum value at x = c.
If f is differentiable at x = c, what must be true of f (c),
and why?
If f is a differentiable function, then the values x =
c at which the sign of the derivative f (x) changes
are the locations of the local extrema of f . Use
this information to find the local extrema of the function f (x) = sin x. Illustrate your answer on a graph of
y = sin x.
If f is a differentiable function, then the values x = c
at which the sign of the second derivative f (x)
changes are the locations of the inflection points of
f . Use this information to find the inflection points of
the function f (x) = sin x. Illustrate your answer on a
graph of y = sin x.
CHAPTER REVIEW, SELF-TEST, AND CAPSTONES
Before you progress to the next chapter, be sure you are familiar with the definitions, concepts, and basic skills outlined here.
The capstone exercises at the end bring together ideas from this chapter and look forward to future chapters.
Definitions
Give precise mathematical definitions or descriptions of each
of the concepts that follow. Then illustrate the definition with
a graph or algebraic example, if possible.
the graphical interpretations of a tangent line and a secant
line to a graph
the real–world interpretations of position, velocity, and
acceleration
the real–world interpretations of average rate of change and
instantaneous rate of change
the formal definition of the derivative of a function f at a
point x = c (both z → x form and h → 0 form)
the formal definition of the derivative of a function f , as a
function (both z → x form and h → 0 form)
the formal definitions of the tangent line and the instantaneous rate of change of the graph of a function f at a point
x=c
what it means for a function f to be differentiable, left differentiable, and right differentiable at a point x = c.
what it means for a function f to be differentiable on an
open or closed interval I
what it means to say that y is an implicit function of x, and
the meaning of implicit differentiation
the definitions of the hyperbolic functions sinh x, cosh x,
and tanh x in terms of exponential functions
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
December 1, 2012
237
Chapter Review, Self-Test, and Capstones
Theorems
Fill in the blanks to complete each of the following theorem
statements:
If a function f is differentiable at x = c, then f is
x = c.
If f and f −1 are inverse functions, then for all appropriate
values of x we can write the derivative of f −1 (x) in terms
.
of the derivative of f (x) as follows:
at
f (x) = kf (x) for some constant k if and only if f is a/an
function.
For all real numbers x, cosh2 x − sinh2 x =
d
(sin x) =
dx
d
(tan x) =
dx
d
(cot x) =
dx
d
(sin−1 x) =
dx
d
(sec−1 x) =
dx
d
(cosh x) =
dx
d
(sinh−1 x) =
dx
d
(tanh−1 x) =
dx
.
Notation and Differentiation Rules
Leibniz notation: Describe the meanings of each of the mathematical expressions that follow. Translate expressions written
in Leibniz notation to “prime” notation, and vice versa.
f (x)
df
dx
dg dt t=3
d 2y
dx 2
f (2)
f (x)
dy
dx
d 2
(x )
dx
d
dx
d
( y(x))
dx
d (x 2 )
dx −1
d 2
(x )
dx
f (5) (x)
Derivatives of basic functions: Fill in the blanks to differentiate
each of the given basic functions. You may assume that k, m,
and b are appropriate constants.
d
(k) =
dx
d
(x) =
dx
d
(mx + b) =
dx
d k
(x ) =
dx
d √
( x) =
dx
d
dx
d x
(e ) =
dx
d kx
(e ) =
dx
d
(ln x) =
dx
1
=
x
d x
(b ) =
dx
d
(logb x) =
dx
d
(cos x) =
dx
d
(sec x) =
dx
d
(csc x) =
dx
d
(tan−1 x) =
dx
d
(sinh x) =
dx
d
(tanh x) =
dx
d
(cosh−1 x) =
dx
Derivatives of combinations: Fill in the blanks to complete each
of the given differentiation rules. You may assume that f and
g are differentiable everywhere.
(kf ) (x) =
( f + g) (x) =
( f − g) (x) =
( fg) (x) =
f
g
(x) =
( f ◦ g) (x) =
d
(ln |x|) =
dx
Skill Certification: Basic Derivatives
Basic definition-of-derivative calculations: Find the derivatives of
the functions that follow, using (a) the h → 0 definition of the
derivative and (b) the z → x definition of the derivative.
1. f (x) = x 2
2. f (x) = x 3
1
x
√
5. f (x) = x
1
x2
1
6. f (x) = √
x
3. f (x) =
4. f (x) =
Calculating basic derivatives: Find the derivatives of the functions that follow, using the differentiation rules developed
in this chapter. (The last two exercises involve hyperbolic
functions.)
1
x 4 − 5x 3 + 2
√
9. f (x) = (x 2x − 1 )−3
7. f (x) =
8. f (x) =
x−2
(x − 1)(x + 3)
10. f (x) = x−1 (4 − x)2
CONFIRMING PAGES
12:17
TKmaster2010
WHF00153/FREE087-Taalman
238
Chapter 2
November 21, 2012
Derivatives
1−
11. f (x) = √ x
x
√
3x 2x + 1
12. f (x) =
1−x
13. f (x) = |x|
14. f (x) = |3x + 1|
15. f (x) = e x sin x
16. f (x) =
17. f (x) = sin(e x )
18. f (x) = e sin x
19. f (x) = ln(tan2 x)
20. f (x) = x 3 sec x
1
21. f (x) =
sin−1 (x 2 )
22. f (x) =
26. f (x) =
x3 + 1
, f (2) = 6
31. f (x) = 8e 4x + 1, f (0) = 3
32. f (x) = 2x sec x 2 tan x 2 , f (0) = 2
ex
sin x
1
, f (0) = 1
1 + 4x
1
34. f (x) =
, f (0) = 1
1 + 4x 2
x
35. f (x) =
, f (0) = 1
1 + 4x 2
1
, f (0) = 0
36. f (x) =
1 − 4x 2
33. f (x) =
sin2 x + cos2 x
csc x
24. f (x) = x x
23. f (x) = x23x+1
25. f (x) = tanh3 (x 5 )
3x 2
30. f (x) = √
sinh−1 x
Differentiating with respect to different variables: Find each
derivative described.
tanh−1 x
37. If 3v2 + xv − 1 = 0, find
Calculating antiderivatives: For each exercise that follows, find
a function f that has the given derivative f and value f (c).
In each case you can find the answer with an educated
guess-and-check process. The last exercise involves an inverse
hyperbolic function.
dv
.
dx
dx
38. If 3v2 + xv − 1 = 0, find .
dv
dA
.
39. If A = πr 2 , find
dr
40. If A = πr 2 , and A and r are both functions of time t,
dA
.
find
dt
27. f (x) = −32, f (0) = 4
dy
.
dt
dy
42. If y = 3x 2 t − t k , where t and k are constant, find .
dx
41. If y = 3x 2 t − t k , where x and k are constant, find
28. f (x) = −32x + 4, f (0) = 100
29. f (x) = x(3x + 1), f (2) = 4
Capstone Problems
A.
The sum rule for differentiation: Use the definition of the
derivative to prove that the derivative of a sum of functions f (x) + g(x) is equal to the sum of their derivatives
f (x) + g (x).
B.
The power rule for differentiation: Use the definition of the
derivative and factoring formulas to prove that for any
positive integer k, the derivative of x k is kx k−1 .
C.
Rates of change from data: The following table lists the
consumption of gasoline in billions of gallons in the
United States from 1994 to 2000:
Year 1994 1995 1996 1997 1998 1999 2000
Gas
109
111
113
117
118
121
122
(a) Compute the average rate of change in gasoline
consumption in the United States for each year
from 1994 to 2000.
(b) During which year was gasoline consumption increasing most rapidly? Least rapidly? Estimate the
instantaneous rates of change in gasoline consumption during those years.
D. Derivatives and graphical behavior: In the next chapter we
will see that we can get a lot of information about the
graph of a function f by looking at the signs of f (x) and
its first and second derivatives. Let’s do this for the function f (x) = x 3 − 3x 2 − 9x + 27.
(a) Find the roots of f , and then determine the intervals
on which f is positive or negative.
(b) Find the roots of f , and then determine the intervals
on which f is positive or negative.
(c) Find the roots of f , and then determine the intervals
on which f is positive or negative.
(d) The graph of f will be above the x-axis when f (x) is
positive and below the x-axis when f (x) is negative.
Moreover, the graph of f will be increasing when f is
positive and decreasing when f is negative. Finally,
the graph of f will be concave up when f is positive and concave down when f is negative. Given
this information and your answers from parts (a)–(c),
sketch a careful, labeled graph of f .
CONFIRMING PAGES
22:0
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
C H A P T E R 3
Applications of the Derivative
3.1
The Mean Value Theorem
The Derivative at a Local Extremum
Rolle’s Theorem
The Mean Value Theorem
Examples and Explorations
3.2
The First Derivative and Curve Sketching
Derivatives and Increasing/Decreasing Functions
Functions with the Same Derivative
The First-Derivative Test
Examples and Explorations
3.3
f
⫹
f
⫹
⫺
Optimization
Finding Global Extrema
Translating Word Problems into Mathematical Problems
Examples and Explorations
w
l
3.5
Related Rates
Related Quantities Have Related Rates
Volumes and Surface Areas of Geometric Objects
Similar Triangles
Examples and Explorations
3.6
f’
The Second Derivative and Curve Sketching
Derivatives and Concavity
Inflection Points
The Second-Derivative Test
Curve-Sketching Strategies
Examples and Explorations
3.4
⫺
r
L’Hôpital’s Rule
Geometrical Motivation for L’Hôpital’s Rule
0
∞
and
0
∞
Using Logarithms for the Indeterminate Forms 00 , 1∞ , and ∞0
L’Hôpital’s Rule for the Indeterminate Forms
Examples and Explorations
Chapter Review, Self-Test, and Capstones
CONFIRMING PAGES
lim
x→c
f (x)
g(x)
f ’’
20:18
TKmaster2010
WHF00153/FREE087-Taalman
240
3.1
Chapter 3
November 24, 2012
Applications of the Derivative
THE MEAN VALUE THEOREM
Local extrema, critical points, and their relationships
Rolle’s Theorem and the Mean Value Theorem
Using critical points to find local extrema
The Derivative at a Local Extremum
Suppose a function f has a local maximum at some point x = c. This means that the value
f (c) is greater than or equal to all other nearby f (x) values. The following definition makes
this notion precise:
DEFINITION 3.1
Local Extrema of a Function
(a) f has a local maximum at x = c if there exists some δ > 0 such that f (c) ≥ f (x) for
all x ∈ (c − δ, c + δ).
(b) f has a local minimum at x = c if there exists some δ > 0 such that f (c) ≤ f (x) for
all x ∈ (c − δ, c + δ).
Intuitively, at a local extremum, the tangent line of a function must be either horizontal or
undefined; for example, consider the following three graphs:
Local maximum with
horizontal tangent line
Local minimum with
horizontal tangent line
y
Local maximum with
no tangent line
y
c
x
y
c
x
c
x
When a function has a horizontal or an undefined tangent line at a point, its derivative
at that point is either zero or undefined. We call such points critical points of the function:
DEFINITION 3.2
Critical Points of a Function
A point x = c in the domain of f is called a critical point of f if f (c) = 0 or f (c) does
not exist.
Notice that only points in the domain of f are considered critical points. For example, con1
1
sider the function f (x) = , whose derivative is f (x) = − 2 . Clearly f (0) does not exist;
x
x
however, since x = 0 is not in the domain of f , it is not called a critical point.
The preceding graphs suggest that every local extremum is also a critical point. This
seemingly obvious relationship between critical points and extrema turns out to be the
foundation on which we will build two more theorems that are key in our development of
calculus:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.1
THEOREM 3.3
241
The Mean Value Theorem
Local Extrema are Critical Points
If x = c is the location of a local extremum of f , then x = c is a critical point of f .
The converse of this theorem is not true. That is, not every critical point is a local extremum
of f ; see Example 1. Although the implication in Theorem 3.3 is intuitively obvious just
by thinking about graphs and the behavior of the derivative at local maxima and minima,
actually proving it requires a somewhat subtle argument. The key is to look at the left and
right derivatives at a local extremum. Theorem 3.3 is also known as Fermat’s Theorem for
Local Extrema, when formulated equivalently as saying that if x = c is a local extremum and
f is differentiable at x = c, then f (c) must be zero.
Proof. We will prove the case for local maxima and leave the similar proof for local minima to
Exercise 64. Suppose x = c is the location of a local maximum of f . If f (c) does not exist, then x = c
is a critical point and we are done. It now suffices to show that if f (c) exists, then it must be equal
to 0. We will do so by examining the right and left derivatives at x = c.
Since x = c is the location of a local maximum of f , there is some δ > 0 such that for all
x ∈ (c − δ, c + δ), f (c) ≥ f (x), and thus f (x) − f (c) ≤ 0. In the case where x ∈ (c, c + δ), it follows that
x > c, which means that x − c is positive. Thus in this case
f + (c) = lim+
x→c
f (x) − f (c)
≤ 0, and therefore
x−c
f (x) − f (c)
≤ 0.
x−c
By a similar argument for x ∈ (c − δ, c), we have x − c < 0 and f (x) − f (c) ≤ 0, and therefore
f − (c) = lim−
x→c
f (x) − f (c)
≥ 0.
x−c
Since we are assuming that f (c) exists, we know that both f + (c) and f − (c) must exist and be equal to
f (c). We have just shown both that f (c) ≤ 0 and that f (c) ≥ 0. Therefore, we must have f (c) = 0,
as desired.
Rolle’s Theorem
Suppose a differentiable function f has two roots x = a and x = b. What can you say about
the graph of f between a and b? The three graphs that follow next provide a clue; if the
graph of a function is smooth and unbroken, then somewhere between each root of f the
function must turn around, and at that turning point it must have a local extremum with a
horizontal tangent line:
y
y
a
c
b
x
y
a c
b
x
a
c b
x
The preceding discussion is a summary of both the statement and the proof of the following
key theorem:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
242
Chapter 3
THEOREM 3.4
November 24, 2012
Applications of the Derivative
Rolle’s Theorem
If f is continuous on [a, b] and differentiable on (a, b), and if f (a) = f (b) = 0, then there
exists at least one value c ∈ (a, b) for which f (c) = 0.
Actually, Rolle’s Theorem also holds in the more general case where f (a) and f (b) are equal
to each other (not necessarily both zero). For example, Rolle’s Theorem is also true if f (a) =
f (b) = 5, or if f (a) = f (b) = −3, and so on, because vertically shifting a function by adding
a constant term does not change its derivative. However, the classic way to state Rolle’s
Theorem is with f (a) and f (b) both equal to zero.
Proof. Rolle’s Theorem is an immediate consequence of the Extreme Value Theorem from
Section 1.4 and the fact that every extremum is a critical point. Suppose f is continuous on the
closed interval [a, b] and differentiable on the open interval (a, b), with f (a) = f (b) = 0. By the Extreme Value Theorem, we know that f attains both a maximum and a minimum value on [a, b]. If
one of these extreme values occurs at a point x = c in the interior (a, b) of the interval, then x = c
is a local extremum of f . By the previous theorem, this means that x = c is a critical point of f .
Since f is assumed to be differentiable at x = c, it follows that f (c) = 0 and we are done.
It remains to consider the special case where all of the maximum and minimum values of
f on [a, b] occur at the endpoints of the interval (i.e., at x = a or at x = b). In this case, since
f (a) = f (b) = 0, the maximum and minimum values of f (x) must both equal zero. For all x in [a, b]
we would have 0 ≤ f (x) ≤ 0, which means that f would have to be the constant function f (x) = 0
on [a, b]. Since the derivative of a constant function is always zero, in this special case we have
f (x) = 0 for all values of c in (a, b), and we are done.
Just as the Intermediate Value Theorem and the Extreme Value Theorem illustrate basic
properties of continuous functions, Rolle’s Theorem illustrates a basic property of functions
that are both continuous and differentiable. Like those two theorems before, Rolle’s Theorem is a theorem about existence, not calculation; it tells you that there must exist some
value c ∈ (a, b) where the derivative of f is zero, but it does not tell you what that value is. It
is important to note that the continuity and the differentiability hypotheses of Rolle’s Theorem are essential: If a function f fails to be continuous on [a, b] or fails to be differentiable
on (a, b), then the conclusion of Rolle’s Theorem does not necessarily follow; see Example 2.
The Mean Value Theorem
The Mean Value Theorem is a generalization of Rolle’s Theorem to the case where f (a) and
f (b) are not necessarily equal. Suppose f is a continuous, differentiable function. What can
we say about the derivative of f between two points x = a and x = b? The three graphs
that follow suggest an answer: Somewhere between a and b the slope of the tangent line
must be the same as the slope of the line from (a, f (a)) to (b, f (b)). If you turn your head so
that the green line is horizontal in each figure, you can see that these figures are similar to
rotated versions of the earlier figures used in illustrating Rolle’s Theorem.
y
y
a
c
b
x
y
a
c
b
x
a
c
b
x
Algebraically, this means that there must be some c ∈ (a, b) whose derivative value f (c)
is equal to the average rate of change of f on [a, b]:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.1
THEOREM 3.5
The Mean Value Theorem
243
The Mean Value Theorem
If f is continuous on [a, b] and differentiable on (a, b), then there exists at least one value
c ∈ (a, b) such that
f (b) − f (a)
f (c) =
.
b−a
The “mean” in the “Mean Value Theorem” refers to an average. Basically, this theorem says
that if a function is continuous on a closed interval and differentiable on its interior, then
there is always at least one place in the interval where the instantaneous rate of change of
the function is equal to its average rate of change over the whole interval. As a real–world
example, suppose you drove at an average speed of 50 miles per hour on a short road trip.
The Mean Value Theorem guarantees that at some point along your journey you must have
been travelling at exactly 50 miles per hour.
The Mean Value Theorem is intuitively clear if you believe that you can just “turn your
head to the side” and see Rolle’s Theorem. In fact, the proof of the Mean Value Theorem is
based on an algebraic version of this intuition:
Proof. Suppose f is a function that is continuous on [a, b] and differentiable on (a, b), and let l(x)
be the secant line from (a, f (a)) to (b, f (b)). The idea of the proof is to “turn our heads” algebraically.
To rotate so that the secant line l(x) plays the role of the x-axis, we will consider the function g(x) =
f (x) − l(x). The graph of this new function g(x) will have roots at x = a and x = b, and we will be
able to apply Rolle’s Theorem.
Since the secant line l(x) has slope
tion is
l(x) =
f (b) − f (a)
and passes through the point (a, f (a)), its equab−a
f (b) − f (a)
(x − a) + f (a).
b−a
This means that the function g(x) = f (x) − l(x) is equal to
g(x) = f (x) −
f (b) − f (a)
(x − a) − f (a).
b−a
If we want to apply Rolle’s Theorem to g(x), then we must first verify that g(x) satisfies all the
hypotheses of Rolle’s Theorem. First, g(x) is continuous on [a, b] because it is a combination of
continuous functions. Second, g(x) is differentiable on (a, b) because f is differentiable on (a, b).
Finally, g(a) = 0 and g(b) = 0:
g(a) = f (a) −
f (b) − f (a)
(a − a) − f (a) = f (a) − 0 − f (a) = 0,
b−a
g(b) = f (b) −
f (b) − f (a)
(b − a) − f (a) = f (b) − ( f (b) − f (a)) − f (a) = 0.
b−a
Since Rolle’s Theorem applies to the function g(x), we can conclude that there exists some
c ∈ (a, b) for which g (c) = 0. How does this conclusion relate to our original problem? To answer
that, we must first calculate g (x):
f (b) − f (a)
d
f (x) −
(x − a) − f (a)
← definition of g(x)
g (x) =
dx
b−a
= f (x) −
f (b) − f (a)
(1) − 0
b−a
←
= f (x) −
f (b) − f (a)
.
b−a
← simplify
f (b) − f (a)
and f (a) are constants
b−a
Rolle’s Theorem now guarantees that there exists a c ∈ (a, b) for which g (c) = 0. By our previous
calculation, for this value of c we have g (c) = f (c) −
as desired.
f (b) − f (a)
f (b) − f (a)
= 0, and therefore f (c) =
,
b−a
b−a
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
244
Chapter 3
November 24, 2012
Applications of the Derivative
Examples and Explorations
EXAMPLE 1
Not every critical point is a local extremum
Show that x = 1 is a critical point of f (x) = x 3 − 3x 2 + 3x. Then use a graph to show that
x = 1 is not a local extremum of f .
SOLUTION
If f (x) = x 3 − 3x 2 + 3x, then f (x) = 3x 2 − 6x + 3 and thus f (1) = 3(1)2 − 6(1) + 3 = 0.
Therefore x = 1 is a critical point of f . However, looking at the following graph of f , we can
see that f has neither a local minimum nor a local maximum at x = 1:
x = 1 is a critical point but not an extremum
y
2
1
1
EXAMPLE 2
2
x
The hypotheses of Rolle’s Theorem are important
Sketch graphs of three functions that fail to satisfy the hypotheses of Rolle’s Theorem, for
which the conclusion of Rolle’s Theorem does not follow.
SOLUTION
The function f in the first graph that follows at the right fails to be differentiable on (1, 3),
and therefore can “turn around” at x = 2 without having a horizontal tangent line. For this
function, there is no value of c ∈ (a, b) with f (c) = 0.
g is not continuous on [1, 3]
f is not differentiable on (1, 3)
y
h is not continuous on [1, 3]
y
y
2
2
2
1
1
1
1
2
3
4
x
1
2
3
4
x
1
2
3
4
x
In the second graph, the function g(x) fails to be continuous at the very place where we
would have expected its derivative to be zero. Since this function is not continuous at x = 2,
it is also not differentiable at x = 2, and there is no value c ∈ (1, 3) with g (c) = 0.
The third graph illustrates a function h(x) that fails to be continuous at the right endpoint x = 3 of the interval. There is no value c ∈ (1, 3) with h (c) = 0; the function never
has to “turn around,” since it just jumps down to the root at x = 3.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.1
EXAMPLE 3
The Mean Value Theorem
245
Applying Rolle’s Theorem
Use Rolle’s Theorem to show that there must exist some value of c in (−2, 5) at which the
function f (x) = x 2 − 3x − 10 has a horizontal tangent line. Then use f to find such a value
c algebraically, and verify your answer with a graph.
SOLUTION
First notice that f (x) = x 2 − 3x − 10 = (x + 2)(x − 5) has roots at x = −2 and at x = 5.
Since f is a polynomial, it is continuous and differentiable. In particular, it is continuous on
[−2, 5] and differentiable on (−2, 5). Therefore Rolle’s Theorem applies to the function f ,
and we can conclude that there must exist some value of c ∈ (−2, 5) for which f (c) = 0. At
this value of c the graph of f will have a horizontal tangent line.
Rolle’s Theorem tells us that there exists some c ∈ (−2, 5) where f (c) = 0, but it
doesn’t tell us exactly where. We can find such a c by solving the equation f (x) = 0. Since
3
f (x) = x 2 − 3x − 10, we have f (x) = 2x − 3, which is equal to zero when x = . Therefore f
2
3
2
has a horizontal tangent line at c = , which is in the interval (−2, 5). The following graph
3
2
illustrates that f (x) = x 2 − 3x − 10 does appear to have a horizontal tangent line at x = .
Horizontal tangent at x =
3
2
y
2 1
2
1
2
3
4
5
x
4
6
8
10
12
14
EXAMPLE 4
Using critical points and Rolle’s Theorem to find local extrema
The function f (x) = x (x − 1)(x − 3) is a cubic polynomial with one local maximum and one
local minimum. Use Rolle’s Theorem to identify intervals on which these extrema exist.
Then use derivatives to find the exact locations of these extrema.
SOLUTION
The roots of f (x) = x (x − 1)(x − 3) are x = 0, x = 1, and x = 3. Since f is a polynomial,
it is continuous and differentiable everywhere. Therefore Rolle’s Theorem applies on the
intervals [0, 1] and [1, 3], and it tells us that at least one critical point must exist inside each
of these intervals.
The critical points of f are the possible locations of the local extrema that we seek. To
find the critical points we must solve the equation f (x) = 0. It is simpler to do some algebra
before differentiating:
f (x) =
d
d
(x (x − 1)(x − 3)) = (x 3 − 4x 2 + 3x) = 3x 2 − 8x + 3.
dx
dx
By the quadratic formula, we have f (x) = 0 at the points
x=
−(−8) ±
√
√
82 − 4(3)(3)
8 ± 28
4± 7
=
=
.
2(3)
6
3
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
246
Chapter 3
November 24, 2012
Applications of the Derivative
These x-values are approximately x ≈ 0.451 and x ≈ 2.215. If we look at the graph of f ,
then we can see that the smaller of these two x-values is the location of the local maximum
and the larger is the location of the local minimum; see the figure that follows.
CHECKING
THE ANSWER
The graph of f (x) = x(x − 1)(x − 3) is shown next. Notice that the local extrema do seem
to occur at the values we just found.
Extrema at x ≈ 0.451 and x ≈ 2.215
4
⫺1
4
⫺8
EXAMPLE 5
Applying the Mean Value Theorem
Use the Mean Value Theorem to show that there is some value c ∈ (0, 2) at which the
tangent line to the function f (x) = x 2 − 2 has slope 2. Then use f to find such a value c
algebraically, and verify your answer with a graph.
SOLUTION
The function f (x) = x 2 − 2 is a polynomial and thus is continuous and differentiable; in
particular it is continuous on [0, 2] and differentiable on (0, 2). Therefore, the Mean Value
Theorem applies to this function on the interval [0, 2]. The slope of the line from (0, f (0))
to (2, f (2)) is
f (2) − f (0)
(22 − 2) − (02 − 2)
2 − (−2)
4
=
=
= = 2.
2−0
2
2
2
By the Mean Value Theorem, there must exist at least one point c ∈ (0, 2) with f (c) = 2.
f
To find such a value of c algebraically, observe that the derivative of f (x) = x 2 − 2 is
= 2x + 0 = 2x. We want to find c ∈ (0, 2) with f (c) = 2, so we solve:
(x)
f (c) = 2 =⇒ 2c = 2 =⇒ c = 1.
The point c = 1 is indeed in the interval (0, 2), and f (1) = 2, so we are done. The following
figure illustrates that f (x) = x 2 − 2 does appear to have the same slope at x = 1 as the
secant line from (0, f (0)) to (2, f (2)).
f (1) equals average rate of change on [0, 2]
y
4
3
2
1
1
1
1
2
x
2
3
4
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.1
TEST YOUR
? UNDERSTANDING
The Mean Value Theorem
247
Is every critical point a local extremum? Is every local extremum a critical point?
What is the role of δ in the definitions of local extrema given in Definition 3.1?
Why do you think f (a) = f (b) would be a sufficient hypothesis in the statement of
Rolle’s Theorem? Think about the situation graphically.
Can Rolle’s Theorem tell you the exact location of a root of f ?
How is the Mean Value Theorem related to Rolle’s Theorem?
EXERCISES 3.1
Thinking Back
Review of definitions and theorems: State each theorem or
definition that follows in precise mathematical language.
Then give an illustrative graph or example, as appropriate.
f has a local maximum at x = c
f has a local minimum at x = c
f is continuous on [a, b]
f is differentiable on (a, b)
The secant line from (a, f (a)) to (b, f (b))
The right derivative f + (c) at a point x = c
The left derivative f − (c) at a point x = c
The Extreme Value Theorem
The Intermediate Value Theorem
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: Rolle’s Theorem is a special case of the
Mean Value Theorem.
(b) True or False: The Mean Value Theorem is so named
because it concerns the average (or “mean”) rate of
change of a function on an interval.
(c) True or False: If f is differentiable on R and has an extremum at x = −2, then f (−2) = 0.
(d) True or False: If f has a critical point at x = 1, then
f has a local minimum or maximum at x = 1.
(e) True or False: If f is any function with f (2) = 0 and
f (8) = 0, then there is some c in the interval (2, 8)
such that f (c) = 0.
(f) True or False: If f is continuous and differentiable on
[−2, 2] with f (−2) = 4 and f (2) = 0, then there is
some c ∈ (−2, 2) with f (c) = −1.
(g) True or False: If f is continuous and differentiable on
[0, 10] with f (5) = 0, then f has a local maximum or
minimum at x = 5.
(h) True or False: If f is continuous and differentiable on
[0, 10] with f (5) = 0, then there are some values a
and b in (0, 10) for which f (a) = 0 and f (b) = 0.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function with a local minimum at x = 3 that is
continuous but not differentiable at x = 3.
(b) A function with a local maximum at x = −2 that is
not differentiable at x = −2 because of a removable
discontinuity.
(c) A function with a local minimum at x = 1 that
is not differentiable at x = 1 because of a jump
discontinuity.
3. If f has a local maximum at x = 1, then what can you say
about f (1)? What if you also know that f is differentiable
at x = 1?
4. If f has a local minimum at x = 0 and a local maximum
at x = 2, what can you say about f (0) and f (2)? Is there
anything else you can say about f ?
5. Suppose that f is defined on (−∞, ∞) and differentiable
everywhere except at x = −2 and x = 4, and that f (x) = 0
only at x = 0 and x = 5. List all the critical points of f and
sketch a possible graph of f .
6. Suppose that f is defined for x = 0 and differentiable
everywhere except at x = 0 and x = 1, and that f (x) = 0
only at x = ±2. List all the critical points of f and sketch
a possible graph of f .
7. If a continuous, differentiable function f has zeroes at
x = −4, x = 1, and x = 2, what can you say about f on [−4, 2]?
8. If a continuous, differentiable function f is equal to 2 at
x = 3 and at x = 5, what can you say about f on [3, 5]?
9. If a continuous, differentiable function f has values
f (−2) = 3 and f (4) = 1, what can you say about f on
[−2, 4]?
10. Restate Theorem 3.3 so that its conclusion has to do with
tangent lines.
11. Restate Rolle’s Theorem so that its conclusion has to do
with tangent lines.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
248
Chapter 3
November 24, 2012
Applications of the Derivative
12. Restate the Mean Value Theorem so that its conclusion
has to do with tangent lines.
In Exercises 13–22, sketch the graph of a function that satisfies
the given description. Label or annotate your graph so that it
is clear that it satisfies each part of the description.
13. A function that satisfies the hypothesis, and therefore the
conclusion, of Rolle’s Theorem on [2, 6].
14. A function that satisfies the hypothesis, and therefore the
conclusion, of the Mean Value Theorem.
15. A function f that satisfies the hypotheses of Rolle’s Theorem on [−2, 2] and for which there are exactly three values
c ∈ (−2, 2) that satisfy the conclusion of the theorem.
16. A function f that satisfies the hypothesis of the Mean
Value Theorem on [0, 4] and for which there are exactly
three values c ∈ (0, 4) that satisfy the conclusion of the
theorem.
17. A function f that is defined on [−2, 2] with f (−2) = f (2) =
0 such that f is continuous everywhere, differentiable everywhere except at x = −1, and fails the conclusion of
Rolle’s Theorem.
18. A function f defined on [1, 5] with f (1) = f (5) = 0 such
that f is continuous everywhere except for x = 2, differentiable everywhere except at x = 2, and fails the conclusion
of Rolle’s Theorem.
19. A function f defined on [−3, −1] with f (−3) = f (−1) = 0
such that f is continuous everywhere except at x = −1
and differentiable everywhere except at x = −1, and fails
the conclusion of Rolle’s Theorem.
20. A function f defined on [0, 4] such that f is continuous everywhere, differentiable everywhere except at x = 2, and
fails the conclusion of the Mean Value Theorem with
a = 0 and b = 4.
21. A function f defined on [−3, 3] such that f is continuous everywhere except at x = 1, differentiable everywhere
except at x = 1, and fails the conclusion of the Mean Value
Theorem with a = −3 and b = 3.
22. A function f defined on [−2, 0] such that f is continuous
everywhere except at x = −2, differentiable everywhere
except at x = −2, and fails the conclusion of the Mean
Value Theorem with a = −2 and b = 0.
Skills
For each graph of f in Exercises 23–26, approximate all the
values x ∈ (0, 4) for which the derivative of f is zero or does
not exist. Indicate whether f has a local maximum, minimum,
or neither at each of these critical points.
y
23.
y
24.
2
3
2
1
2
3
4
x
2
1
1
1
2
34. f (x) = 21−ln x
36. f (x) = sec x
For each graph of f in Exercises 37–40, explain why f satisfies the hypotheses of Rolle’s Theorem on the given interval
[a, b]. Then approximate any values c ∈ (a, b) that satisfy the
conclusion of Rolle’s Theorem.
[a, b] = [−3, 1]
37.
1
1
ln 2x
x
35. f (x) = cos x
33. f (x) =
y
3
2
2
1
[a, b] = [−3, 3]
38.
y
x
2
3
2
y
25.
1
4
1
3 2 1
1
y
26.
4
3
2
3
1
1
1
2
3
4
x
2
1
1
1
2
3
4
x
Find the critical points of each function f in Exercises 27–36.
Then use a graphing utility to determine whether f has a
local minimum, a local maximum, or neither at each of these
critical points.
27. f (x) = (x − 1.7)(x + 3)
28. f (x) = x 3 + x 2 + 1
29. f (x) = 3x 4 + 8x 3 − 18x 2
30. f (x) = (2x − 1)5
31. f (x) = 3x − 2e x
32. f (x) = 3 x − 2 x
y
4
2
3
1
2
3
x
[a, b] = [−1, 1]
40.
3
1
3
3
y
1
2
2
x
[a, b] = [0, 4]
39.
2
2
1
1
2
3
4
2
x
1
3 2 1
1
1
2
3
x
Determine whether or not each function f in Exercises 41–48
satisfies the hypotheses of Rolle’s Theorem on the given
interval [a, b]. For those that do, use derivatives and
algebra to find the exact values of all c ∈ (a, b) that satisfy the
conclusion of Rolle’s Theorem.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.1
41. f (x) = x 3 − 4x 2 + 3x, [a, b] = [0, 3]
42. f (x) = x 3 − 4x 2 + 3x, [a, b] = [1, 3]
43. f (x) = x 4 − 3.24x 2 − 3.04, [a, b] = [−2, 2]
x 2 − 4x
, [a, b] = [0, 4]
44. f (x) = 2
x − 4x + 3 [a, b] = [−3, 0]
y
51.
π 3π
2 2
46. f (x) = sin 2x, [a, b] = [0, 2π ]
47. f (x) = e x (x 2 − 2x), [a, b] = [0, 2]
√ √ 48. f (x) = ln |x 2 − 1|, [a, b] = − 2, 2
4
[a, b] = [0, 2]
y
2
1
2
1
1
2
3
x
3 2 1
1 1
1
1
x
1
1
2
3
4
x
2
Determine whether or not each function f in Exercises 53–60
satisfies the hypotheses of the Mean Value Theorem on the
given interval [a, b]. For those that do, use derivatives and algebra to find the exact values of all c ∈ (a, b) that satisfy the
conclusion of the Mean Value Theorem.
54. f (x) = x 3 − 2x + 1, [a, b] = [0, 6]
55. f (x) = −x 3 + 3x 2 − 7, [a, b] = [−2, 3]
8
7
6
5
4
3
2
1
3
2
3
2
53. f (x) = x 2 + , [a, b] = [−3, 2]
y
4
4
1
x
[a, b] = [−1, 3]
50.
y
2
For each graph of f in Exercises 49–52, explain why f satisfies
the hypotheses of the Mean Value Theorem on the given interval [a, b] and approximate any values c ∈ (a, b) that satisfy
the conclusion of the Mean Value Theorem.
49.
3
[a, b] = [0, 4]
52.
6
5
4
3
2
1
45. f (x) = cos x, [a, b] = − ,
249
The Mean Value Theorem
56. f (x) = (x 2 − 1)(x 2 − 4), [a, b] = [−3, 3]
57. f (x) = ln(x 2 + 1), [a, b] = [0, 1]
58. f (x) = 2 x , [a, b] = [0, 3]
1
2
3
x
59. f (x) = sin x, [a, b] = 0,
π
2
60. f (x) = tan x, [a, b] = [−π, π ]
Applications
61. The cost of manufacturing a container for frozen orange
juice is C(h) = h2 − 7.4h + 13.7 cents, where h is the
height of the container in inches. Your boss claims that
the containers will be cheapest to make if they are 4
inches tall. Use Theorem 3.3 to quickly show that he is
wrong.
62. Last night at 6 p.m., Linda got up from her blue easy chair.
She did not return to her easy chair until she sat down
again at 8 p.m. Let s(t) be the distance between Linda and
her easy chair t minutes after 6 p.m. last night.
(a) Sketch a possible graph of s(t), and describe what
Linda did between 6 p.m. and 8 p.m. according to
your graph. (Questions to think about: Will Linda
necessarily move in a continuous and differentiable
way? What are good ranges for t and s?)
(b) Use Rolle’s Theorem to show that at some point
between 6 p.m. and 8 p.m., Linda’s velocity v(t) with
respect to the easy chair was zero. Find such a place
on the graph of s(t).
63. It took Alina half an hour to drive to the grocery store that
is 20 miles from her house.
(a) Use the Mean Value Theorem to show that, at some
point during her trip, Alina must have been travelling
exactly 40 miles per hour.
(b) Why does what you have shown in part (a) make
sense in real-world terms?
Proofs
64. Prove the part of Theorem 3.3 that was not proved in the
reading: If a function f has a local minimum at x = c, then
either f (c) does not exist or f (c) = 0.
65. Prove Rolle’s Theorem: If f is continuous on [a, b] and differentiable on (a, b), and if f (a) = f (b) = 0, then there is
some value c ∈ (a, b) with f (c) = 0.
66. Prove the Mean Value Theorem: If f is continuous on
[a, b] and differentiable on (a, b), then there is some value
c ∈ (a, b) with f (c) =
f (b) − f (a)
.
b−a
68. Follow the method of proof that we used for Rolle’s
Theorem to prove the following slightly more general
theorem: If f is continuous on [a, b] and differentiable on
(a, b), and if f (a) = f (b), then there is some value c ∈ (a, b)
with f (c) = 0.
69. Use Rolle’s Theorem to prove the slightly more general
theorem from Exercise 68: If f is continuous on [a, b] and
differentiable on (a, b), and if f (a) = f (b), then there is
some value c ∈ (a, b) with f (c) = 0. (Hint: Apply Rolle’s
Theorem to the function g(x) = f (x) − f (a).)
67. Use Rolle’s Theorem to prove that if f is continuous and
differentiable everywhere and has three roots, then its
derivative f has at least two roots.
CONFIRMING PAGES
20:18
WHF00153/FREE087-Taalman
250
Chapter 3
November 24, 2012
Applications of the Derivative
Thinking Forward
Sign analyses for second derivatives: Repeat the instructions of
the previous block of problems, except find sign intervals for
the second derivative f instead of the first derivative.
Sign analyses for derivatives: For each function f that follows,
find the derivative f . Then determine the intervals on which
the derivative f is positive and the intervals on which the
derivative f is negative. Record your answers on a sign chart
for f , with tick-marks only at the x-values where f is zero or
undefined.
f (x) =
f (x) =
3.2
x2
x
+1
sin x
ex
f (x) = x 2 3 x
f (x) =
x
x2 + 1
f (x) = x 2 3 x
f (x) =
sin x
ex
f (x) = ln( ln x)
f (x) = ln( ln x)
THE FIRST DERIVATIVE AND CURVE SKETCHING
The relationship between the derivative and increasing/decreasing functions
Proving that all antiderivatives of a function differ by a constant
Using the first-derivative test to determine whether critical points are maxima, minima, or neither
Derivatives and Increasing/Decreasing Functions
In Section 0.4 we defined a function f to be increasing on an interval if, for all a and b in
the interval with b > a, f (b) > f (a). In other words, the height of f at points farther to the
right are higher. Similarly, f is decreasing on an interval if, for all b > a in the interval,
f (b) < f (a). These definitions can be difficult to work with if we wish to find the intervals
on which a given function is increasing or decreasing. Luckily, the derivative will provide
us with an easier method.
We have seen that the first derivative f in some sense measures the direction of the
graph of a function f at each point, since f is the associated slope function for f . In particular, if f is positive at a point x = c, then the graph of f must be moving in an upwards
direction, that is, increasing, as it passes x = c. Similarly, if f (c) is negative, then the graph
of f must be decreasing at x = c. For example, we can divide the real-number line into intervals according to where the function f (x) = x 3 −3x 2 −9x+11 is increasing or decreasing,
as shown in the figure that follows at the left. This same division into subintervals describes
where the derivative f (x) = 3x 2 − 6x − 9 is positive and negative, as shown at the right.
Intervals where f increases/decreases
Intervals where f is positive/negative
y
de
cre
asi
ng
4 3 2 1
1
2
incre
asing
y
in
cre
as
in
g
TKmaster2010
3
4
5
positive
6
4 3 2 1
positive
1
2
3
4
5
negative
(, 1]
[1, 3]
[3, )
(, 1]
[1, 3]
CONFIRMING PAGES
[3, )
6
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
The First Derivative and Curve Sketching
251
This relationship between intervals on which f is increasing or decreasing exactly when
f is positive or negative, respectively, holds in general. The only wrinkle is at the extremum
of a function f ; for example, in the graph at the left we say that f is increasing on (−∞, −1],
which includes the extremum at x = −1, because it is true that for all b > a less than or
equal to x = −1 we do have f (b) > f (a). However, the derivative at x = −1 is not positive,
but zero. Therefore f is positive only on the interior (−∞, −1) of the interval, but the
function f is increasing on the entire interval (−∞, −1].
THEOREM 3.6
The Derivative Measures Where a Function is Increasing or Decreasing
Let f be a function that is differentiable on an interval I.
(a) If f is positive in the interior of I, then f is increasing on I.
(b) If f is negative in the interior of I, then f is decreasing on I.
(c) If f is zero in the interior of I, then f is constant on I.
Theorem 3.6 is intuitively obvious if we consider the slopes of tangent lines on increasing and decreasing graphs. To prove this theorem formally we require the Mean Value
Theorem.
Proof. We’ll prove part (a) here and leave the similar proofs of parts (b) and (c) to Exercises 89
and 90, respectively. The key to the proof will be the Mean Value Theorem.
Let f be a function that is differentiable on an interval I and whose derivative f is positive on
the interior of that interval. Suppose also that a, b ∈ I with b > a. By the definition of increasing,
we must show that f (b) > f (a). Since f is differentiable, and thus also continuous, on the interval I,
and since [a, b] is contained in the interval I, f satisfies the hypotheses of the Mean Value Theorem
on [a, b]. Therefore we can conclude that there exists some c ∈ (a, b) such that
f (c) =
f (b) − f (a)
.
b−a
To show that f (b) > f (a) it suffices to show that f (b) − f (a) > 0; with this in mind we can rearrange
the preceding equation as
f (b) − f (a) = f (c) (b − a).
Since c ∈ (a, b), it follows that c is in the interior of I, and thus by hypothesis f (c) > 0. Furthermore,
since b > a, we have (b − a) > 0. Therefore f (b) − f (a) is the product of two positive numbers and
must itself be positive, which is what we wanted to show.
Up to this point we could only graphically approximate the intervals on which a function
is increasing or decreasing. With Theorem 3.6 we can now find these intervals algebraically,
by examining the sign of f . We have thus reduced the difficult problem of finding the intervals on which a function is increasing or decreasing to the much simpler problem of finding
the intervals on which an associated function—the derivative—is positive or negative.
Functions with the Same Derivative
If two functions differ by a constant, then obviously they will have the same derivative,
because the derivative of a constant is zero. For example, f (x) = x 3 and g(x) = x 3 + 10 differ
by a constant because their difference g(x) − f (x) is equal to the constant 10, and their
derivatives f (x) = 3x 2 and g (x) = 3x 2 + 0 = 3x 2 are equal.
Although it is less obvious, the converse is also true: Any two functions that have the
same derivative must differ by a constant. Algebraically, this means that if you find one
CONFIRMING PAGES
20:18
WHF00153/FREE087-Taalman
252
Chapter 3
November 24, 2012
Applications of the Derivative
antiderivative of a function, then all other antiderivatives of that function differ from the
one that you found by a constant. For example, one antiderivative of 3x 2 is x 3 , and thus all
antiderivatives of 3x 2 are of the form x 3 + C for some constant C. Graphically, this means
that if two functions have the same derivative, then one is a vertical shift of the other. For
example, the graph of 3x 2 is shown next at the left and five of its antiderivatives x 3 + C
are shown at the right, for C = 0, C = ±10, and C = ±20. The red graph of y = 3x 2
yields information about every one of the blue graphs y = x 3 + C, regardless of the vertical
shift C.
Graph of y = 3x 2
All antiderivatives of 3x 2 are of the form x 3 + C
y
y
40
30
20
20
3 2 1
10
1
2
3
x
20
3 2 1
THEOREM 3.7
1
2
3
x
40
Functions with the Same Derivative Differ by a Constant
If f (x) = g (x) for all x ∈ [a, b], then, for some constant C, f (x) = g(x) + C for all
x ∈ [a, b].
Proof. Suppose f (x) = g (x) for all x ∈ [a, b]. Then f (x) − g (x) = 0, which by the difference
d
rule means that ( f (x) − g(x)) = 0 on [a, b]. By the third part of Theorem 3.6 this means that
dx
the function f (x) − g(x) is constant for all x ∈ [a, b], say, f (x) − g(x) = C for some real number C.
Therefore f (x) = g(x) + C for all x ∈ [a, b].
The First-Derivative Test
In the previous section we saw that the set of critical points of a function, that is, the values
of x for which f is zero or does not exist—is a complete list of all the possible local extrema
of f . We now develop a method for using the first derivative to determine which critical
points are local maxima, which are local minima, and which are neither.
Suppose f (c) = 0 and that f is differentiable near c. Then if f is not constant, there are
four different ways that f can behave near x = c:
y
c
x
not an
extremum
c
not an
extremum
x
CONFIRMING PAGES
c
ng
easi
decr
ing
reas
dec
c
minimum
x
y
incre
asing
g
easin
decr
maximum
f is decreasing
on both sides of x = c
incre
asing
y
f is increasing
on both sides of x = c
ng
easi
decr
y
f changes from decreasing
to increasing at x = c
incre
asing
f changes from increasing
to decreasing at x = c
incr
easi
ng
TKmaster2010
x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
253
The First Derivative and Curve Sketching
On the one hand, if f changes from increasing to decreasing at x = c, then f has a local
maximum at x = c. If f changes from decreasing to increasing at the point x = c, then f has
a local minimum at x = c. On the other hand, if f does not change direction at x = c, then
f does not have a local extremum at x = c. Since a function f is increasing or decreasing
according to whether its derivative is positive or negative, we can record the preceding four
situations on sign charts for f , with the corresponding information for f recorded as the
following sign analyses:
f changes from positive
to negative at x = c
max
CAUTION
f
c
f changes from negative
to positive at x = c
f’
⫹
c
f’
c
f is negative
on both sides of x = c
f
neither
f
min
⫺
f is positive
on both sides of x = c
f’
f
neither
c
f’
Please recall that, throughout this book, we will make tick-marks on sign charts only at the
locations where the function is zero or fails to exist. We will mark with “DNE” any locations
where the function does not exist, but we will not explicitly mark the zeroes. You should
assume that any unlabeled tick-marks on a sign chart are zeroes of the function in question.
Do not make additional tick-marks on your sign charts if you follow this convention.
Using a sign chart for the first derivative f to test whether critical points of f are local
maxima, minima, or neither is quite sensibly known as the first-derivative test. The
statement of the test is wordy and its proof is somewhat technical, but its meaning is equivalent to information obtained from the previous sign chart analyses.
THEOREM 3.8
The First-Derivative Test
Suppose x = c is the location of a critical point of a function f , and let (a, b) be an open
interval around c that is contained in the domain of f and does not contain any other
critical points of f . If f is continuous on (a, b) and differentiable at every point of (a, b)
except possibly at x = c, then the following statements hold.
(a) If f (x) is positive for x ∈ (a, c) and negative for x ∈ (c, b), then f has a local maximum at x = c.
(b) If f (x) is negative for x ∈ (a, c) and positive for x ∈ (c, b), then f has a local minimum at x = c.
(c) If f (x) is positive for both x ∈ (a, c) and x ∈ (c, b), then f does not have a local
extremum at x = c.
(d) If f (x) is negative for both x ∈ (a, c) and x ∈ (c, b), then f does not have a local
extremum at x = c.
Proof. We will prove parts (a) and (c) here and leave the similar proofs of parts (b) and (d) to
Exercises 91 and 92, respectively. The proof will be an application of Theorem 3.6. Suppose x = c
is a critical point of f , and let (a, b) be an interval around x = c satisfying the hypotheses of the
theorem.
To prove part (a), suppose f (x) > 0 for x ∈ (a, c) and f (x) < 0 for x ∈ (c, b), that is, suppose
that f is increasing on (a, c] and decreasing on [c, b). We will show that f (c) ≥ f (x) for all x ∈ (a, b),
which will tell us that f has a local maximum at x = c. Given any x ∈ (a, b), there are three cases to
consider. First, if x = c, then clearly f (c) = f (x). Second, if a < x < c, then since f is increasing on
(a, c], we have f (x) < f (c). Third, if c < x < b, then since f is decreasing on [c, b), we have f (c) > f (x).
In all three cases we have f (c) ≥ f (x), and therefore f has a local maximum at x = c.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
254
Chapter 3
November 24, 2012
Applications of the Derivative
To prove part (c), suppose that f (x) > 0 for all x ∈ (a, c) ∪ (c, b). Then by Theorem 3.6, f must
be increasing on all of (a, b). Now, the point x = c cannot be the location of a local maximum of
f , because, for all x > c in (a, b), f (x) > f (c), since f is increasing on (a, b). But neither can x = c
be a local minimum of f , because, for all x < c in (a, b), f (x) < f (c). Therefore f has neither a local
minimum nor a local maximum at x = c.
Now, to find the local extrema of a function f , we need only find the critical points of
f and then test each one with the first-derivative test. In other words, we find the derivative f , determine where f is zero or does not exist, and then make a sign chart for f around these critical points to determine whether f has a local maximum, a local minimum,
or neither at each critical point. This method will find all local extrema for functions f that
are defined on open intervals. For functions defined on closed or half-closed intervals, we
will also have to consider endpoint extrema, which we will discuss in Section 3.4.
Examples and Explorations
EXAMPLE 1
Using the derivative to determine where a function is increasing or decreasing
Use Theorem 3.6 to determine the intervals on which each of the following functions are
increasing or decreasing:
(a) f (x) = x 3
(b) g(x) = x 2 − 2x + 1
(c) h(x) =
x 2 + 10x + 1
x−2
SOLUTION
(a) If f (x) = x 3 , then f (x) = 3x 2 , which is positive as long as x = 0. Therefore f is positive
on (−∞, 0) and (0, ∞). By Theorem 3.6 we can say that f is increasing on the entire
half-closed intervals (−∞, 0] and [0, ∞). Thus f is increasing on all of R. Notice that,
as shown in the graph that follows, the tangent line to f (x) = x 3 is horizontal at x = 0.
However, the function f (x) = x 3 is still increasing everywhere: For all real numbers
b > a we have b3 > a3 , even if one of a or b is zero.
f (x) = x 3 is increasing everywhere
y
8
4
2
1
1
2
x
4
8
(b) If g(x) = x 2 − 2x + 1, then g (x) = 2x − 2. Recall that functions can change signs only at
roots, non-domain points, and discontinuities. The derivative g (x) = 2x − 2 is always
defined and continuous, and g (x) = 0 when 2x − 2 = 0 (i.e., when x = 1). Therefore we need only check the sign of g (x) to the left and right of x = 1. For example,
g (0) = 2(0) − 2 = −2 is negative, and g (2) = 2(2) − 2 = 2 is positive, as the following
sign chart shows:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
g’(0) ⫽ ⫺2 < 0,
so g’(x) < 0 for
all x in (⫺∞, 1)
The First Derivative and Curve Sketching
tick-mark
means
g’(1) ⫽ 0
⫺
255
g’(2) ⫽ 2 > 0,
so g’(x) > 0 for
all x in (1, ∞)
⫹
g’
1
By Theorem 3.6, the information on the sign chart for g (x) shows that g(x) is
decreasing on (−∞, 1] and increasing on [1, ∞). However, from this point forward we
will record the closed-endpoint information only when we need it and instead record
only the open-interval information. Therefore here we would say simply that g(x) is
decreasing on (−∞, 1) and increasing on (1, ∞).
(c) By the quotient rule, the derivative of h(x) is
(2x + 10)(x − 2) − (x 2 + 10x + 1)(1)
(x − 2)2
2
x − 4x − 21
(x + 3)(x − 7)
=
=
.
(x − 2)2
(x − 2)2
h (x) =
Notice that we simplified the preceding equation so that it would be immediately clear
that h (x) = 0 when x = −3 and x = 7 and that h (x) does not exist when x = 2, a
point that is not in the domain of h(x) in the first place. To determine the sign chart
for h (x) we need only check the sign of h (x) one time between each of these critical
points and non-domain points. For example, using the factored form of h (x) we have:
(−)(−)
(positive)
(−)2
(+)(−)
h (5) :
(negative)
(+)2
(+)(−)
(negative)
(−)2
(+)(+)
h (10) :
(postiive)
(+)2
h (−5) :
h (0) :
We can record this information on a sign chart for h (x) as follows, with annotations
above the chart for the corresponding increasing/decreasing behavior of h(x):
h
DNE 3
2
h’
7
Reading off the sign chart, we see that h(x) is increasing on (−∞, −3) and (7, ∞) and
decreasing on (−3, 2) and on (2, 7).
CHECKING
THE ANSWER
To check our answers to parts (b) and (c) of the example, we can simply graph g(x) and h(x).
Notice that the graph of g(x) shown next at the left does appear to be decreasing on (−∞, 1)
and increasing on (1, ∞). The graph of h(x) at the right does appear to be increasing on
(−∞, −3) and (7, ∞) and decreasing elsewhere.
g(x) = x 2 − 2x + 1
h(x) =
4
x 2 + 10x + 1
x−2
40
⫺10
⫺1
15
3
0
⫺40
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
256
Chapter 3
EXAMPLE 2
November 24, 2012
Applications of the Derivative
Interpreting the sign chart for the derivative of a function
Let f be a function whose derivative f has the signs indicated on the sign chart shown.
Recall our convention that tick-marks on this sign chart indicate the locations where f is
zero (if unlabeled) or does not exist (if marked with “DNE”). Use the sign chart to sketch
possible graphs for f and f .
⫹
⫺
DNE
⫺1
⫹
f’
2
SOLUTION
This sign chart says that f is positive on the intervals (−∞, −1) and (2, ∞), and negative
on the interval (−1, 2). It also says that f (−1) = 0 and that f (2) does not exist. Given this
information, one possible sketch of f is shown next at the left.
We can determine a lot about the graph of a function f from the graph of its derivative
f . The function f must be increasing on (−∞, −1) and (2, ∞), and decreasing on (−1, 2),
with a horizontal tangent line at x = −1, and a non-differentiable point at x = 2. One
possible sketch of a function f that has these characteristics is shown next at the right.
Notice that the information about f does not tell us how high or low to sketch the graph
of f . In fact, any vertical shift of f would do just as well, since every antiderivative of f is of
the form f (x) + C.
Possible graph of f Possible graph of f
y
y
1
2
x
1
2
x
EXAMPLE 3
Detailed curve-sketching analyses using the first derivative
Sketch the graph of each of the given functions. Along the way, determine all local extrema
and important features of the graph.
(a) f (x) = x 3 − 3x + 2
(b) f (x) =
x2
x 2 − 2x + 1
SOLUTION
(a) For f (x) = x 3 − 3x + 2 we have f (x) = 3x 2 − 3. This derivative is always defined and
continuous, so the critical points of f are just the places where f (x) = 0:
3x 2 − 3 = 0 =⇒ 3x 2 = 3 =⇒ x 2 = 1 =⇒ x = ±1.
These critical points divide the real-number line into three intervals, namely,
(−∞, −1), (−1, 1), and (1, ∞). We’ll test the sign of f on each interval by testing the
sign of f (x) at one point in that interval, say, at x = −2, x = 0, and x = 2:
f (−2) = 3(−2)2 − 3 = 12 − 3 = 9 > 0,
f (0) = 3(0)2 − 3 = 0 − 3 = −3 < 0,
f (2) = 3(2)2 − 3 = 12 − 3 = 9 > 0.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
The First Derivative and Curve Sketching
257
Recording this information on a sign chart for f , with consequences for f illustrated
above the chart, we have
max
f
min
1
f’
1
The maximum and minimum of f listed on the chart were identified by the firstderivative test: Since f changes from positive to negative at x = −1, f changes from
increasing to decreasing there; thus the point x = −1 must be a local maximum of f .
Since f changes from negative to positive at x = 1, f changes from decreasing to
increasing there; thus f must have a local minimum at the point x = 1.
With all the first-derivative information we have collected, we can sketch a reasonable graph of f by plotting just a few points. As a general rule it is a good idea to
plot the points (c, f (c)) for each critical point x = c and then use the information about
the derivative to connect the dots accordingly. In this example, the critical points are
x = −1 and x = 1, so we calculate:
f (−1) = (−1)3 − 3(−1) + 2 = −1 + 3 + 2 = 4,
f (1) = (1)3 − 3(1) + 2 = 1 − 3 + 2 = 0.
Using this information, we can make a sketch of the graph of f (x) = x 3 − 3x + 2, as
shown next. Compare the features of this graph with the information in the sign chart
for f :
y
5
4
3
2
1
3 2 1
1
1
2
3
x
(b) Before we do anything else, we should note that the domain of
x2
x2
=
x 2 − 2x + 1
(x − 1)2
is all points except x = 1. The function f is zero only at x = 0, and at all other points in
its domain it is positive, since both its numerator and denominator are perfect squares.
So far we know that the entire graph of f will be above the x-axis, except at the root
(0, 0) and at the non-domain point x = 1, where something interesting may occur.
f (x) =
To determine local extrema and increasing/decreasing behavior, we must find the
derivative of f (x) =
x2
x2
. Because we will be interested in where f − 2x + 1
is zero or does
not exist, we also simplify as much as possible:
f (x) =
(2x)(x 2 − 2x + 1) − (x 2 )(2x − 2)
−2x(x − 1)
−2x
=
=
.
(x 2 − 2x + 1)2
(x − 1)4
(x − 1)3
This derivative is zero when its numerator is zero but its denominator is not, that is,
when x = 0. The derivative is undefined when x = 1, which makes sense because
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
258
Chapter 3
November 24, 2012
Applications of the Derivative
the original function f is also undefined at x = 1. The critical point x = 0 and the
non-domain point x = 1 are the only places at which the sign of f can change.
Now we make a sign chart representing the sign of the derivative f between each
of these critical points. For example, we can find that f (−1) < 0, f (0.5) > 0, and
f (2) < 0:
min
f
DNE
0
f’
1
We can identify the local minimum at x = 0 indicated on the chart with the firstderivative test, since f (x) changes sign from negative to positive at x = 0. However,
the same type of argument cannot be applied to the point x = 1: Even though f (x)
changes sign at x = 1, the function f does not have a local extremum at x = 1, because
it is not defined at that point.
Information about limits can be important to figuring out the key features of a
x2
graph, so we will examine limits of f (x) = 2
at any “interesting” places. In this
x − 2x + 1
case the interesting limits to consider are at the point x = 1, where the function does
not exist, and at the ends as x → ±∞. As x → 1 we have x 2 → 1 and x 2 − 2x + 1 =
1
(x − 1)2 → 0+ , and therefore the limit in question is of the form + . Thus
0
lim
x→1
x2
x 2 − 2x + 1
= ∞.
We can calculate the limit of f as x → ∞ with the familiar strategy of dividing
numerator and denominator by the highest power of x:
1
x2
1
1/x 2
=
lim 2
= 1.
= lim
2
1
x→∞ x − 2x + 1 1/x 2
x→∞
1−0+0
1− + 2
x
A similar argument shows that
x2
lim
x→−∞ x 2 − 2x + 1
x
is also equal to 1. (We could also ap-
peal to Theorem 1.32 and notice that the numerator and denominator of the function
in question are polynomials of the same degree, and thus that the limit of the quo1
tient as x → ±∞ is the ratio = 1 of the leading coefficients of those polynomials.)
1
Since lim f (x) = ∞, we can see that f has a vertical asymptote at the non-domain
x→1
point x = 1, with the function approaching ∞ on both sides of this asymptote. Since
lim f (x) = 1 and lim f (x) = 1, f has a horizontal asymptote on both the left and
x→∞
the right at y = 1.
x→−∞
Putting all of this information together, we get the following graph:
y
6
5
4
3
2
1
4
2
2
4
6
x
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
TEST YOUR
? UNDERSTANDING
The First Derivative and Curve Sketching
259
If f is positive on (0, 3), why does it make sense that we can say that f is increasing on
all of [0, 3]?
Can you give an example of a function f for which f is zero on all of (−1, 1)?
If x 3 − 4x 2 − 2 is one antiderivative of some function f , what can you say about the
other antiderivatives of f ?
If f is defined, continuous, and differentiable at x = 3, and if f changes sign at x = 3,
then what can you say about f at x = 3?
Can you sketch an example of a function f with a critical point at x = 2 but no local
extremum at x = 2?
EXERCISES 3.2
Thinking Back
Differentiation: Find the derivative of each function f , and then
simplify as much as possible.
(2x − 1)3
(3x + 1)2
f (x) = (2x − 1)3 (3x + 1)2
f (x) =
f (x) = 3x 2 e −4x
f (x) = sin( ln x)
Solving equations: For each of the following functions g(x), find
the solutions of g(x) = 0 and also find the values of x for which
g(x) does not exist.
√
1 −1/2
x
(1 + 5x) − x(5)
2
g(x) =
(1 + 5x)2
√
g(x) = 2x x − 1 + x 2
g(x) =
1
(x − 1)−1/2
2
e x (1 − e x ) − e x (−e x )
(1 − e x )2
Sign analyses: For each of the following functions g(x), use
algebra and a sign chart to find the intervals on which g(x)
is positive and the intervals on which g(x) is negative.
g(x) = 6x 2 − 18x
g(x) =
(x + 2)(x − 1)4
ex
g(x) = 2 − x − 2x 2 + x 3
g(x) =
3x 2 − 5x − 2
sin2 x
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If f (x) < 0 for all x ∈ (0, 3), then f is
decreasing on [0, 3].
(b) True or False: If f is increasing on (−2, 2), then f (x) ≥
0 for all x ∈ (−2, 2).
(c) True or False: If f (x) = 2x, then f (x) = x 2 +C for some
constant C.
(d) True or False: If f is continuous on (1, 8) and f (3) is
negative, then f is negative on all of (1, 8).
(e) True or False: If f changes sign at x = 3, then
f (3) = 0.
(f) True or False: If f (−2) = 0, then f has either a local
maximum or a local minimum at x = −2.
(g) True or False: If x = 1 is the only critical point of f and
f (0) is positive, then f (2) must be negative.
(h) True or False: If f (1) is negative and f (3) is positive,
then f has a local minimum at x = 2.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function that is decreasing on (−∞, 0), increasing
on (0, ∞), and undefined at x = 0.
(b) A function that is decreasing on (−∞, 0] and increasing on [0, ∞).
(c) A function that is always positive and always decreasing, on all of R.
3. State the definition of what it means for a function f to be
increasing on an interval I and what it means for a function f to be decreasing on an interval I.
4. Describe what a critical point is, intuitively and in mathematical language. Then describe what a local extremum
is. How are these two concepts related?
5. Can a point x = c be both a local extremum and a critical point of a differentiable function f ? Both an inflection
point and a critical point? Both an inflection point and a
local extremum? Sketch examples, or explain why such a
point cannot exist.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
260
Chapter 3
November 24, 2012
Applications of the Derivative
6. Suppose f is a function that may be non-differentiable at
some points. Can a point x = c be both a local extremum
and a critical point of such a function f ? Both an inflection
point and a critical point? Both an inflection point and a
local extremum? Sketch examples, or explain why such a
point cannot exist.
14. Suppose f is a function that is continuous and differentiable everywhere and that the derivative of f is
7. Suppose f is defined and continuous everywhere. Why is
testing the sign of the derivative f at just one point sufficient to determine the sign of f on the whole interval
between critical points of f ?
8. Describe what the first-derivative test is for and how
to use it. Sketch graphs and sign charts to illustrate your
description.
15. If g(x) and h(x) are both antiderivatives of some function f (x), then what can you say about the function
g(x) − h(x)?
9. Sketch the graph of a function f with the following properties:
f (x) =
(x − 1)(x − 2)
.
x −3
What are the critical points of f ?
16. If g(x) is an antiderivative of f (x), then what is the relationship between the functions g(x) + 10 and f (x)?
17. One of the graphs shown is a function f and the other is
its derivative f . Which one is which, and why?
f is continuous and defined on R;
f (0) = 5;
Graph I
Graph II
y
y
5
5
f (−2) = −3 and f (−2) = 0;
f (1) does not exist;
f is positive only on (−2, 1).
2 1
1
2
4
3
x
2 1
1
2
4
3
x
10. Sketch the graph of a function f with the following properties:
f is continuous and defined on R;
5
f has critical points at x = −3, 0, and 5;
5
18. One of the graphs shown is a function f and the other is
its derivative f . Which one is which, and why?
f has inflection points at x = −3, −1, and 2.
11. Use the definitions of increasing and decreasing to argue
that f (x) = x 4 is decreasing on (−∞, 0] and increasing on
[0, ∞). Then use derivatives to argue the same thing.
12. Use the definition of increasing to argue that f (x) = x 5 is
increasing on all of R. Then use derivatives to argue the
same thing.
Graph I
Graph II
y
y
1
13. Suppose f is a function that is continuous and differentiable everywhere and that the derivative of f is
f (x) = 1 + x 2 − 4.
1
2 1
1
2
4
3
x
2 1
1
What are the critical points of f ?
1
2
4
3
x
1
Skills
For each function f graphed in Exercises 19–22, sketch a
possible graph of its derivative f .
19.
y
y
21.
2
y
20.
1
1
2
3
2
x
2
3
1
1
2
1
4 3 2 1
1
2
3
y
2
1
1
3
1
22.
3
1
2
3
x
1
1
2
x
CONFIRMING PAGES
2
3
x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
Each graph in Exercises 23–26 represents the derivative f of some function f . Use the given graph of f to sketch a possible graph of f .
y
23.
y
24.
4
3
51.
2
0
1
1
1
For each sign chart for f in Exercises 51–56, sketch possible graphs of both f and f . On each sign chart, unlabeled tick-marks are locations where f (x) is zero and
x-values where f (x) does not exist are indicated by tick-marks
labeled “DNE.”
3
2
2
1
1
2
x
2
2
1
1
1
x
3
2
4
53.
y
26.
54.
3
2
1
2
3
2
55.
2
3
8
f’
f’
1
DNE DNE 56.
4 3 2 1
1
2
3
x
Use a sign chart for f to determine the intervals on which each
function f in Exercises 27–38 is increasing or decreasing. Then
verify your algebraic answers with graphs from a calculator or
graphing utility.
27. f (x) = 2x 3 − 9x 2 + 1
DNE 1
1
2
0
f’
3
x
2
1
f’
2
3
y
f’
5
3
25.
52.
2
3
261
The First Derivative and Curve Sketching
28. f (x) = x 3 + 4x 2 + 4x − 5
1
4
f’
Sketch careful, labeled graphs of each function f in
Exercises 57–82 by hand, without consulting a calculator or
graphing utility. As part of your work, make sign charts for
the signs, roots, and undefined points of f and f , and examine any relevant limits so that you can describe all key points
and behaviors of f .
3x + 1
x2 − 1
57. f (x) = (x − 2)(3x + 1)
58. f (x) = x 2 − x + 100
59. f (x) = x 3 − x 2 − x + 1
60. f (x) = x3 − 9x + 1
31. f (x) = e x (x − 2)
ex
32. f (x) =
1 + ex
61. f (x) = x 3 − 6x 2 + 12x
62. f (x) = x(x 2 − 4)
33. f (x) = ln(x 2 + 1)
34. f (x) = ln( ln x)
63. f (x) = (2x + 11)(x 2 + 10)
64. f (x) = 3x 5 − 10x 4
29. f (x) =
x
x2 + 4
35. f (x) = sin
30. f (x) =
π
x
2
65. f (x) = (1 − x 4 )7
36. f (x) = cos (x)
2
38. f (x) =
37. f (x) = sin x cos x
x+1
x−1
√
69. f (x) = x 2 + 1
67. f (x) =
1
sin x
Use the first-derivative test to determine the local extrema of
each function f in Exercises 39–50. Then verify your algebraic
answers with graphs from a calculator or graphing utility.
71. f (x) =
(x − 1)2
x2 + x − 6
39. f (x) = (x − 2)2 (1 + x)
73. f (x) =
x 2 (x − 1)
(x − 2)2
1 + x + x2
x2 + x − 2
1
43. f (x) =
3 − 2e x
41. f (x) =
45. f (x) = cos(π(x + 1))
47. f (x) = arctan x
49. f (x) = sin(cos
−1
40. f (x) = x 2 (x − 1)(x + 1)
42. f (x) =
(x − 1)2
x+2
44. f (x) = e x (x 2 − x − 1)
77. f (x) = x 2 3 x
46. f (x) = cos(πx)
79. f (x) =
48. f (x) = sin
x)
75. f (x) = x ln x
−1
x
2
50. f (x) = cos(sin
−1
ln x
x
81. f (x) = e x
3
−3x 2 +2x
√
1
x− √
x
1
1
68. f (x) = + 2
x
x
1
70. f (x) = √
x2 − 1
(x − 1)2
72. f (x) = 2
x −1
66. f (x) =
x3
− 3x + 2
2x
76. f (x) =
1 − 2x
74. f (x) =
x2
78. f (x) = x 3 e x
80. f (x) = ln((x−1)(x−2))
82. f (x) = e ln e
x)
CONFIRMING PAGES
x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
262
Chapter 3
November 24, 2012
Applications of the Derivative
Applications
83. Dr. Alina is interested in the behavior of rats trapped in
a long tunnel. Her rat Bubbles is released from the lefthand side of the tunnel and runs back and forth in the
tunnel for 4 minutes. Bubbles’ velocity v(t), in feet per
minute, is given by the following graph.
For Exercises 85 and 86, suppose that Annie is planning a
kayak trip around Orcas Island in August. The tides create
strong currents in several places on the coast of that island.
tide is going out
Velocity of rat in a long tunnel
tide is coming in
kayak
v
kayak
5
t, hours
4
3
c
2
2
1
⫺1
1
2
3
4
t
positive
current
1
⫺2
1
(a) On which time intervals is Bubbles moving towards
the right-hand side of the tunnel?
(b) At which point in time is Bubbles farthest away from
the left-hand side of the tunnel, and why? Do you
think that Bubbles ever comes back to the left-hand
side of the tunnel?
(c) On which time intervals does Bubbles have a positive
acceleration?
(d) Find an interval on which Bubbles has a negative
velocity but a positive acceleration. Describe what
Bubbles is doing during this period.
84. Calvin uses a slingshot to launch an orange straight up in
the air to see what will happen. The distance in feet between the orange and the ground after t seconds is given
by the equation s(t) = −16t 2 + 90t + 5. Use this equation
to answer the following questions:
(a) What is the initial height of the orange? What is the
initial velocity of the orange? What is the initial acceleration of the orange?
(b) What is the maximum height of the orange?
(c) When will the orange hit the ground?
⫺1
⫺2
2
3
4
5
6
7
8
9
10
11
12
t
negative
current
85. Annie has information from the gauge at Point Lawrence to
help her decide the best time to round that point. The tidal
current velocity (in nautical miles per hour) around Point
Lawrence this August can be modeled very simply as
c(t) = 0.86 cos (0.51t + 2.04).
A positive sign on the current indicates that it is going roughly north (the tide is coming in), while a negative sign indicates southward motion of the current. The
time t is given in hours after midnight of the morning
of August 1.
(a) When is c(t) equal to zero? What is the significance
of these times for the tides?
(b) When do the high tides occur? When are the low
tides?
(c) Suppose Annie does not want to have to fight the
current. Approximately when would be good time intervals for Annie to lead her party southward around
the point?
86. After a bad experience on one trip, Annie models the tidal
current velocity around Point Lawrence in a more accurate way as
0.51t
+ 2.13 .
C(t) = 0.90 cos (0.51t + 2.02) + 0.49 cos
2
(a) Plot this function C(t) together with the function c(t)
from the previous exercise on the same axes.
(b) The maximum currents given by c(t) occurred when
t = −4 ± 12.32n hours, for n an integer. Demonstrate
that C(t) does not have maxima at these points.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.2
263
The First Derivative and Curve Sketching
Proofs
87. Prove that every nonconstant linear function is either
always increasing or always decreasing.
constant on I. (Hint: use the Mean Value Theorem to show
that any two numbers a and b in I must be equal.)
88. Prove that every quadratic function has exactly one local
extremum.
91. Prove part (b) of Theorem 3.8: With hypotheses as stated
in the theorem, if x = c is a critical point of f , where
f (x) < 0 to the left of c and f (x) > 0 to the right of c,
then f has a local minimum at x = c .
92. Prove part (d) of Theorem 3.8: With hypotheses as stated
in the theorem, if x = c is a critical point of f , where
f (x) < 0 to the left and to the right of c, then x = c is
not a local extremum of f .
89. Prove part (b) of Theorem 3.6: Suppose f is differentiable
on an interval I; if f is negative on the interior of I, then
f is decreasing on I.
90. Prove part (c) of Theorem 3.6: Suppose f is differentiable
on an interval I; if f is zero on the interior of I, then f is
Thinking Forward
Second-derivative graphs: The three graphs shown are
graphs of a function f and its first and second derivatives f and f , in no particular order. Identify which
graph is which.
Graph I
y
Graph II
y
2
2
1
1
More second-derivative graphs: The three graphs
shown are graphs of a function f and its first and second derivatives f and f , in no particular order. Identify which graph is which.
Graph I
Graph II
y
y
3
3
2
2
1
⫺3 ⫺2 ⫺1
x
1
2
⫺3 ⫺2 ⫺1
3
⫺1
⫺1
⫺2
⫺2
Graph III
y
⫺2
2
3
⫺3 ⫺2 ⫺1
⫺1
1
2
3
⫺3 ⫺2 ⫺1
⫺1
⫺2
⫺2
⫺3
⫺3
3
2
1
⫺1
1
1
x
Graph III
y
2
⫺3 ⫺2 ⫺1
x
1
x
1
2
3
⫺3 ⫺2 ⫺1
⫺1
1
2
⫺2
⫺3
CONFIRMING PAGES
3
x
1
2
3
x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
264
3.3
Chapter 3
November 24, 2012
Applications of the Derivative
THE SECOND DERIVATIVE AND CURVE SKETCHING
Using first and second derivatives to define and detect concavity
The behavior of the first and second derivatives at inflection points
Using the second-derivative test to determine whether critical points are maxima, minima, or neither
Derivatives and Concavity
In Section 0.4 we gave an informal definition of concavity: The graph of a function is
concave up if it “curves upward” and concave down if it “curves downward.” This is equivalent to saying that the graph of a concave-up function lies below its secant lines and above
its tangent lines, and the graph of a concave-down function lies above its secant lines and
below its tangent lines. Now that we know about derivatives, we are finally able to give a
more precise definition of concavity.
DEFINITION 3.9
Formally Defining Concavity
Suppose f and f are both differentiable on an interval I.
(a) f is concave up on I if f is increasing on I.
(b) f is concave down on I if f is decreasing on I.
How does this formal definition of concavity correspond with our intuitive notion of concavity? Consider the functions graphed next. On each graph four slopes are illustrated and
estimated. Notice that when f is concave up, its slopes increase from left to right, and when
f is concave down, its slopes decrease from left to right.
Slopes increase when f is concave up
y
Slopes decrease when f is concave down
y
1 1
3
4
1
3
4
1
x
x
As we have already seen, a function increases where its derivative is positive. Taking this
up one level, we see that the derivative function f is increasing where its derivative function
f is positive. Therefore we can check whether a function f is concave up or concave down
by looking at the sign of its second derivative:
THEOREM 3.10
The Second Derivative Determines Concavity
Suppose both f and f are differentiable on an interval I.
(a) If f is positive on I, then f is concave up on I.
(b) If f is negative on I, then f is concave down on I.
CONFIRMING PAGES
20:18
WHF00153/FREE087-Taalman
November 24, 2012
3.3
265
The Second Derivative and Curve Sketching
Proof. We will prove part (a) and leave part (b) to Exercise 91. Suppose that f and f are differentiable on I and that f (x) > 0 for all x in I. Then since the derivative of f is f , it follows from
Theorem 3.6 that f is increasing on I. By the definition of concavity this means that f is concave
up on the interval I.
For example, we can divide the real-number line into intervals according to where the
function f (x) = x 3 is concave up or concave down. This same division into subintervals
describes where the derivative f (x) = 3x 2 is increasing or decreasing and where the second
derivative f (x) = 6x is positive or negative:
f decreasing, then increasing
f concave down, then up
y
f negative, then positive
y
y
positive
1
1
own
ed
2
ing
rea
s
ng
x
2
con
cav
2
c
inc
av
si
rea
eu
p
dec
con
1
1
2
x
2
1
1
2
x
negative
(, 0]
[0, )
(, 0]
[0, )
(, 0]
[0, )
Inflection Points
Recall from Section 0.1 that the inflection points of a function f are the points in the domain of f at which its concavity changes. Since the sign of f measures the concavity of f ,
we can find inflection points by looking for the places where f changes sign. For example,
if f (x) = x 3 , then f (x) = 6x, which is zero only when x = 0. The sign of f changes from
negative to positive at x = 0, and therefore the concavity of f changes from down to up at
that inflection point.
If x = c is an inflection point of f and f (c) = 0, then the graph of f could look one of
the following four ways near x = c, depending on how f changes concavity and whether f
is increasing or decreasing:
y
y
cav
con
e
up
c
up
co n
x
wn
e do
av
conc
n
dow
conc
ave
n
dow
own
ve d
ca
av
con
c
y
co n c
e up
y
conc
ave
TKmaster2010
co n c
x
ca
ave
c
x
c
ve u
p
x
We know that at an inflection point of a function f , the function changes concavity
and the second derivative f changes sign. What happens to the first derivative f ? The
answer lies in the fact that f is the derivative of f . Thus we know that If f is positive,
then f is increasing, and If f is negative, then f is decreasing. The four possible scenarios
corresponding to the preceding graphs are shown here:
CONFIRMING PAGES
20:18
WHF00153/FREE087-Taalman
266
Chapter 3
November 24, 2012
Applications of the Derivative
IP
f
min
f’
c
f ’’
IP
f
max
f’
c
f ’’
IP
f
max
f’
min
f ’’
c
f
IP
f’
c
f ’’
Notice that in each case the sign of f changes at x = c, causing f to have an inflection point at x = c and in addition causing f to have a local maximum or minimum at
x = c. If you sketch tangent lines on the four graphs shown, you should be able to see
that the slopes are at their maximum or minimum values at the inflection points. For example, in the first graph, the slopes start out large and positive, decrease to a minimum of
zero at x = c, and then increase to larger and larger positive slopes as we move from left
to right.
The Second-Derivative Test
In the previous section we saw how to apply the first-derivative test to determine whether
the critical points of a function were local maxima, local minima, or neither. We can also
use the second derivative to test critical points, by examining the concavity of the function
at each critical point.
Suppose f is a differentiable function and x = c is a critical point of f with f (c) = 0.
Then there are four possible ways that f can behave near x = c, as shown in the figures
that follow. In each case we can examine the second derivative at the point x = c:
y
c
not an
extremum
con
x
conc.
do
wn
c
x
own
c. d
x
c. up
c
a ve
up
c.
n
ow
co
nc
minimum
c
y
y
not an
extremum
cave d
f (c) = 0, f (c) = 0
co n
maximum
up
y
f (c) = 0, f (c) = 0
f (c) = 0, f (c) ≥ 0
on
f (c) = 0, f (c) ≤ 0
con
TKmaster2010
c
x
In the first graph, the second derivative is negative at x = c, so f curves downwards and
has a local maximum at that point. In the second graph, the second derivative is positive
at x = c, so f curves upwards and has a local minimum at that point.
In general, knowing that f (c) = 0 does not tell us whether f has a maximum, minimum, or neither at x = c. The reason is that f (c) = 0 is a possibility in all four of the
cases just graphed. It is obvious that the last two graphs must have f (c) = 0 because in
those cases the function changes concavity at x = c. Perhaps less obvious is that we could
have f (c) = 0 in the first two graphs, where no inflection points occur; for example see
the function f (x) = x 4 that we will examine later in part (a) of Example 1.
What we have just illustrated with the preceding four graphs is the second-derivative
test. When we applied the first-derivative test to a critical point, we were interested in the
sign of f (x) to the left and the right of the critical point. When we use the second derivative
to test a critical point, we will look at the sign of the second derivative at the critical point,
as follows:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
THEOREM 3.11
The Second Derivative and Curve Sketching
267
The Second-Derivative Test
Suppose x = c is the location of a critical point of a function f with f (c) = 0, and suppose
both f and f are differentiable and f is continuous on an interval around x = c.
(a) If f (c) is positive, then f has a local minimum at x = c.
(b) If f (c) is negative, then f has a local maximum at x = c.
(c) If f (c) = 0, then this test says nothing about whether or not f has an extremum
at x = c.
Proof. Suppose both f and f are differentiable in a neighborhood of a critical point x = c with
f (c) = 0. To prove part (a), suppose f (c) > 0. Since f is assumed to be continuous near c, f must be positive in a small neighborhood (c − δ, c + δ) of c. Because f is the derivative of f , it
follows from Theorem 3.6 that f is increasing on (c − δ, c + δ). Since f is increasing near c, and is
zero at c, we must have f (x) < 0 to the immediate left of c and f (x) > 0 to the immediate right
of c. Therefore, by the first-derivative test, f has a local minimum at x = c. The proof of part (b) is
similar and is left to Exercise 92.
To prove part (c), it suffices to exhibit three functions with f (c) = 0 at some point c where
one function has a local maximum at x = c, one function has a local minimum at x = c, and one
function does not have a local extremum at x = c. You will do this in Exercise 10.
Curve-Sketching Strategies
When we graph a function by hand, a good place to start is to use algebra, derivatives, and
sign charts to determine the intervals where the function is positive or negative, increasing
or decreasing, and concave up or down, as well as the coordinates of any roots, extrema,
and inflection points of f . Armed with this information and occasionally a few strategic
function values and limits, we can often sketch a fairly accurate graph.
Of course, actually sketching a graph based on the information just described can
take a little bit of practice. One useful thing to notice is that a continuous, differentiable
function can change sign only at it roots, change direction only at its local extrema, and
change concavity only at its inflection points. This means that on the intervals between
such points the graph of the function is relatively homogeneous. In fact, if f is a sufficiently
well-behaved function, then we can connect each adjacent pair of dots determined by the
coordinates of its local extrema and inflection points with one of the following four types
of arcs:
Increasing and
concave up
Increasing and
concave down
Decreasing and
concave up
Decreasing and
concave down
Here is a summary of what you might consider when sketching the graph of a function
f by hand:
Find the domain of f .
Determine the roots of f and the intervals on which f is positive or negative. Record
this information on a sign chart for f .
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
268
Chapter 3
November 24, 2012
Applications of the Derivative
Find f and determine the points where f is zero or does not exist. Each of these
points that is in the domain of f is a critical point of f and therefore a possible local
extremum of f .
Determine the intervals on which f is positive or negative. These are the same as the
intervals on which f is increasing or decreasing, respectively. Record this information
on a sign chart for f .
Determine whether f has a local minimum, local maximum, or neither at each
critical point.
Find f and determine the points where f is zero or does not exist. Each of these
points that is in the domain of f is a critical point of f and therefore a possible
inflection point of f .
Determine the intervals on which f is positive or negative. These are the same as
the intervals on which f is concave up or concave down, respectively. Record this
information on a sign chart for f .
Determine whether or not each critical point of f is an inflection point of f .
At each local extremum or inflection point x = c, determine the value of f (c). Plot
these key points (c, f (c)).
At each non-continuous or non-differentiable point x = c in the domain of f , deter-
mine the value of f (c). Use limits of f and f to determine the type of discontinuity
or non-differentiable point. Plot these key points (c, f (c)).
Calculate limits of f at any non-domain points and as x → ±∞. Determine any
horizontal or vertical asymptotes and the long-term behavior of the graph.
Between key points and non-domain points, use the sign charts for f and f to
determine which of the four types of arc shapes the graph will have.
Graph the function by connecting the key points with arcs.
For some functions you will not be able to obtain all of the information you want, but in
every case you should explore as much as you can until you can confidently determine the
behavior of the graph of f .
Why would we want to spend time using derivatives and algebra to sketch the graph
of a function by hand when we could make essentially the same sketch with a graphing
calculator in just a few keystrokes? One reason is that graphing by hand provides more
specific information, such as the exact locations and values of the key points on the graph.
Moreover, if we do the work by hand, then we can determine a graphing window that
captures all of the important features of the graph and possibly also discern features of
the graph that are not apparent with a graphing calculator, such as holes and asymptotes.
Derivatives, limits, and algebra not only enable us to sketch graphs by hand, but also help
us correctly interpret graphs made by calculators.
Examples and Explorations
EXAMPLE 1
Using the second derivative to determine concavity and inflection points
Use second derivatives to determine the intervals on which each of the functions that follow
are concave up or concave down. Then determine any inflection points.
(a) f (x) = x 4
(b) g(x) = x 3 − 3x + 2
SOLUTION
(a) We start by calculating the first and second derivatives of f :
f (x) = x 4
=⇒
f (x) = 4x 3
=⇒
f (x) = 12x 2 .
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
269
The Second Derivative and Curve Sketching
Therefore f (x) = 0 only when x = 0, and f is positive to both the left and right of
x = 0, as shown in the following sign chart:
not
f
IP
f ’’
0
Since the sign of f (x) = 12x 2 does not change at x = 0, the function f (x) = x 4 does
not have an inflection point at x = 0. Therefore f (x) = x 4 has no inflection points and
is concave up on both (−∞, 0) and (0, ∞).
(b) Again we start by calculating derivatives:
g(x) = x 3 − 3x + 2
=⇒
g (x) = 3x 2 − 3
=⇒
g (x) = 6x.
Therefore g (x) = 0 only when x = 0. Testing the sign of g (x) to the left and right of
x = 0, we have g (−1) = −6 < 0 and g (1) = 6 > 0. This information about g (x) is
summarized in the following sign chart:
g
IP
g ’’
0
Thus g is concave down on (−∞, 0), is concave up on (0, ∞), and has an inflection
point at x = 0.
CHECKING
THE ANSWER
To verify the preceding calculations, we sketch graphs of f and g. In the graph that follows at
the left we see that the curve f (x) = x 4 is flat enough to have zero curvature at x = 0, but has
positive curvature to both the left and the right of the origin. In the graph at the right we see
that g(x) = x 3 − 3x + 2 does have an inflection point at x = 0, where its concavity changes
from concave-down to concave-up. Note that g(x) is the same function we examined in
Example 3(a) of the previous section.
g(x) has an inflection point at x = 0
y = x 4 has no inflection points
6
15
3
2
3
2
2
0
EXAMPLE 2
Comparing the first- and second-derivative tests
Determine the local extrema of the function f (x) = x 3 − x 2 + 2, using
(a) the first-derivative test
(b) the second-derivative test
SOLUTION
(a) The derivative of f (x) = x 3 − x 2 + 2 is f (x) = 3x 2 − 2x = x(3x − 2).
This derivative f is zero at the points x = 0 and x =
only critical points of f are x = 0 and x =
2
.
3
2
,
3
and always exists. Thus the
To apply the first-derivative test we must
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
270
Chapter 3
November 24, 2012
Applications of the Derivative
find the sign of the derivative between each of these points. For example, we could
calculate:
f (−1) = 5 > 0,
1
2
f
1
4
and f (1) = 1 > 0.
= − < 0,
The following sign chart summarizes this information about f :
max
f
min
0
f’
2
3
By the first-derivative test, f has a local maximum at x = 0 and a local minimum at
2
x= .
3
(b) The calculation for the second derivative starts out the same way, by computing the
2
derivative f (x) = 3x 2 − 2x and then finding the critical points x = 0 and x = . The
3
difference is that we will test the sign of the second derivative f (x) = 6x − 2 at each
of these critical points:
f (0) = −2 < 0
2
3
and f = 2 > 0.
By the second-derivative test, since f is concave down at the critical point x = 0, f has
2
a local maximum at x = 0. Similarly, since f is concave up at the critical point x = ,
3
2
we know that f has a local minimum at x = . This is of course the same conclusion
3
we reached when we applied the first-derivative test.
EXAMPLE 3
A detailed curve-sketching analysis
Use derivatives, algebra, and sign charts to sketch the graph of f (x) = x 5 − 15x 3 . Identify
the coordinates of each root, local extremum, and inflection point.
SOLUTION
We’ll start by finding and then simplifying the functions f , f , and f . We will factor each
function as much as possible so that we can easily identify its roots:
√
√
f (x) = x 5 − 15x 3 = x 3 (x 2 − 15) = x 3 (x − 15 )(x + 15 );
f (x) = 5x 4 − 45x 2 = 5x 2 (x 2 − 9) = 5x 2 (x − 3)(x + 3);
√
√
f (x) = 20x 3 − 90x = 10x(2x 2 − 9) = 10x( 2x − 3)( 2x + 3).
√
From these factorizations we can see that f has roots at x = 0 and x = ± 15 ≈ ±3.87,
that the only possible local extrema of f are x = 0 and x = ±3, and that the only possible
3
inflection points of f are x = 0 and x = ± √ ≈ ±2.12. By checking the signs of f (x), f (x),
2
and f (x) at appropriate values, we obtain the following sign charts:
15
0
3
0
3
2
15
3
0
3
2
f
f’
f ’’
From this information we can see that f has a maximum at x = −3, a minimum at x = 3,
3
and inflection points at x = 0 and x = ± √ . We can also identify the intervals on which
2
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
271
The Second Derivative and Curve Sketching
f is positive or negative, increasing or decreasing, and concave up or concave down, as
summarized in the following three sign charts:
pos
neg
15
pos
neg
0
15
max
f
f
3
0
IP
f
min
3
IP
f’
3
2
0
f
IP
3
2
f ’’
By evaluating f (x) = x 5 − 15x 3 at its key points we can obtain the coordinates of its
roots, extrema, and inflection points:
√
√
roots at (− 15, 0), (0, 0), and ( 15, 0);
local maximum at (−3, 162), local minimum at (3, −162);
3
3
inflection points at − √ , 100.232 , (0, 0), and √ , −100.232 .
2
2
Now we need only plot these points and connect the dots with appropriate arcs according
to the sign information in the sign charts. For example, we can see from the sign charts for
3
f , f , and f that between x = −3 and x = − √ the graph of f should be positive, decreas2
ing, and concave down.
We already have all of the information we need about the derivatives, but for some
people it helps to collect all this information in one place. The arc shapes on each subinterval between key points are recorded on the combined number-line chart that follows.
Note that on this number line the tick-marks represent the locations of all interesting points
on the graph of f , meaning that they are a composite of the tick-marks from the three sign
charts for f , f , and f .
15 3 3
3
2
0
2
3
f
15
Although at this point the shape of the graph is pretty clear, for completeness we should
compute limits at any interesting points. The function f (x) = x 5 −15x 3 has no non-domain
points, discontinuities, or non-differentiable points, so the only limits to check are those
as x → ±∞. We can use what we know about the behavior of fifth-degree polynomials, or
we can just compute these limits directly:
lim (x 5 − 15x 3 ) = lim (x 3 )(x 2 − 15) = ∞,
x→∞
x→∞
lim (x 5 − 15x 3 ) = lim (x 3 )(x 2 − 15) = −∞.
x→−∞
x→−∞
This information tells us that the graph has no horizontal asymptotes and indicates what
happens at the “ends” of the graph of f .
Putting all of the information together into a labeled graph, we have
f (x) = x 5 − 15x 3
y
162.000
100.232
15 3 3
2
3
2
3
15
x
100.232
162.000
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
272
Chapter 3
CHECKING
THE ANSWER
November 24, 2012
Applications of the Derivative
We can verify the graph we constructed in the preceding example by using a calculator or
graphing utility to graph f (x) = x 5 − 15x 3 in a similar graphing window. Notice that we
did indeed capture the major features of this graph. In addition, because we did the work
by hand, we know the exact values of every key point on the graph of f .
Calculator graph of f (x) = x 5 − 15x 3
200
4.5
4.5
200
EXAMPLE 4
A curve-sketching analysis with asymptotes
Sketch an accurate, labeled graph of the function f (x) =
analyses of f , f , and f , and calculate any relevant limits.
6
.
4 − 2x
Include complete sign
SOLUTION
Let’s begin by finding and simplifying the first and second derivatives of f (x). The first
derivative of f (x) = 6(4 − 2 x )−1 is
f (x) = 6(−1)(4 − 2 x )−2 (−( ln 2)2 x ) =
6( ln 2)2 x
.
(4 − 2 x )2
Differentiating that result and then simplifying as much as possible so that we can easily
identify roots, we find that the second derivative of f is
f (x) =
=
6( ln 2)( ln 2)2 x (4 − 2 x )2 − (6( ln 2)2 x )(2)(4 − 2 x )1 (−( ln 2)2 x )
(4 − 2 x )4
6( ln 2)2 2 x (4 − 2 x )((4 − 2 x ) + 2 · 2 x )
6( ln 2)2 2 x (4 + 2 x )
=
.
x
4
(4 − 2 x )3
(4 − 2 )
To determine the intervals on which f , f , and f are positive and negative we must
first locate the values of x for which these functions are zero or do not exist. The function
6
f (x) =
is never zero, but is undefined where its denominator is zero, at x = 2. Looking
x
4−2
at the formula for f (x), we can easily see that it is never zero (since 2 x is never zero), but
is undefined if x = 2 (since the denominator 4 − 2 x is zero for x = 2). Thus x = 2 is the
only critical point of f and the only point we will mark on the number line for f . Similarly,
from the formula for f (x), it is clear that f (x) is never zero (since neither 2 x nor 4 + 2 x
can ever be zero), but is undefined at x = 2. Checking signs on either side of x = 2 for each
function, we obtain the following set of number lines:
2
2
2
f
f’
f ’’
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
The Second Derivative and Curve Sketching
273
We now know that f (x) is positive, increasing, and concave up on (−∞, 2) and negative,
increasing, and concave down on (2, ∞). The graph has no roots, no extrema, and no
inflection points.
It remains to calculate any interesting limits. Since the domain of f is (−∞, 2) ∪ (2, ∞),
6
we must investigate the limits of f (x) =
as x → ±∞ and as x → 2 from the left and
x
4−2
the right. As x → ∞ the denominator 4 − 2 x of f (x) approaches −∞, and thus
6
lim
= 0.
x→∞ 4 − 2 x
As x → −∞, the denominator 4 − 2 x approaches 4, and thus
6
lim
=
x→−∞ 4 − 2 x
6
3
= .
4
2
This information tells us that the graph of f has two horizontal asymptotes: at y = 0 on the
3
right and at y = on the left.
2
As x → 2− we have 4 − 2 x → 0+ and thus
6
lim
x→2− 4 − 2 x
= ∞,
and as x → 2+ we have 4 − 2 x → 0− and thus
lim
6
x→2+ 4 − 2 x
= −∞.
Thus the graph of f has a vertical asymptote at x = 2, where the graph approaches ∞ to
the left of 2 and approaches −∞ to the right of 2.
Putting all of this information together, we can now sketch the graph:
f (x) =
6
4 − 2x
y
1.5
2
x
TEST YOUR
? UNDERSTANDING
Why could we not give a precise mathematical definition of concavity before this section?
The domain points x = c where f (c) = 0 or where f (c) does not exist are the critical
points of the function f . Why?
Why is it not clear to say a sentence such as “Because it is positive, it is concave up”?
How could this information be conveyed more precisely?
Why does it make sense that f is increasing when f is positive?
Suppose x = c is a critical point with f (c) = 0. Why does it make graphical sense that
f has a local minimum at x = c when f is concave up in a neighborhood around x = c?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
274
Chapter 3
November 24, 2012
Applications of the Derivative
EXERCISES 3.3
Thinking Back
Finding the second derivative: For each of the following functions f , calculate and simplify the second derivative f .
f (x) =
x−1
3x 2 − 4x + 2
f (x) = √
f (x) =
1−x
ex
f (x) = e 3x ln(x 2 + 1)
Solving for zeroes and non-domain points: For each of the following expressions, find all values of x for which g(x) is zero
or does not exist.
x
x2 + 1
g(x) =
3x 2 − x − 2
x 4 + 2x 2 − 3
g(x) =
3x + 1
1
−
x−2
x+2
g(x) =
sin x
cos x
g(x) =
e 3x (x − 1)
ln x
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If f (2) = 0, then x = 2 is an inflection
point of f .
(b) True or False: If f is concave up on an interval I, then
it is positive on I.
(c) True or False: If f is concave up on an interval I, then
f is positive on I.
(d) True or False: If f (2) does not exist and x = 2 is in
the domain of f , then x = 2 is a critical point of the
function f .
(e) True or False: If f has an inflection point at x = 3 and
f is differentiable at x = 3, then the derivative f has
a local minimum or maximum at x = 3.
(f) True or False: If f (1) = 0 and f (1) = −2, then f has
a local minimum at x = 1.
(g) True or False: The second-derivative test involves
checking the sign of the second derivative on each
side of every critical point.
(h) True or False: The second-derivative test always
produces exactly the same information as the
first-derivative test.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function f for which f is positive
everywhere, f (x) > 0 for x < −2, and f (x) < 0 for
x > −2.
(b) The graph of a function f for which f (3) = 0,
f (3) = 0, and f (3) = 0.
(c) The graph of a function f for which f (x) is zero at
x = −1, x = 2, and x = 4; f (x) is zero at x = −1,
x = 1, and x = 3; and f (x) is zero at x = 0 and
x = 2.
3. Sketch the graph of a function f that is concave up
everywhere. Then draw five tangent lines on the graph,
and explain how you can see that the derivative of f is
increasing.
4. Sketch the graph of a function f that is concave down
everywhere. Then draw five tangent lines on the graph, and
explain how you can see that the derivative of f is decreasing.
5. State the converse of Theorem 3.10(a). Is the converse
true? If so, explain why; if not, provide a counterexample.
6. State the contrapositive of Theorem 3.10(a). Is the
contrapositive true? If so, explain why; if not, provide a
counterexample.
7. Sketch the graph of a function f that has an inflection
point at x = c in such a way that the derivative f has a
local maximum at x = c. Add tangent lines to your sketch
to illustrate that f does have a local maximum at x = c.
8. Sketch the graph of a function f that has an inflection
point at x = c in such a way that the derivative f has a
local minimum at x = c. Add tangent lines to your sketch
to illustrate that f does have a local minimum at x = c.
9. Show that for f (x) = x 6 we have f (0) = 0 but the point
x = 0 is not an inflection point of f .
10. In this problem we will verify part (c) of Theorem 3.11.
(a) For f (x) = x 3 , show that f (0) = 0 and f (0) = 0
while f does not have a local extremum at x = 0.
(b) For g(x) = x 4 , show that g (0) = 0 and g (0) = 0
while g has a local minimum at x = 0.
(c) For h(x) = −x 4 , show that h (0) = 0 and h (0) = 0
while h has a local maximum at x = 0.
(d) Explain how the parts (a)–(c) show that the secondderivative test does not tell us anything when the
second derivative at a critical point is zero.
11. We could use part (c) from Theorem 3.6 to add a third part
to Theorem 3.10 that would tell us what it means when
f is zero in the interior of an interval I. Fill in the blank
on I.
accordingly: If f is zero on I, then f is
12. Describe what the second-derivative test is for and how
to use it. Sketch graphs and sign charts to illustrate your
description.
13. If a function f has four critical points, how many calculations after finding derivatives are required in order to apply the first-derivative test? The second-derivative test?
14. Describe in words, and then illustrate in pictures, the four
types of arcs that are the building blocks of most continuous, differentiable graphs.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
275
The Second Derivative and Curve Sketching
For Exercises 15–20,sketch the graph of a function f that has the
indicated characteristics. If a graph is not possible, explain why.
17. f negative, f positive, and f positive on [0, 3].
18. f positive, f negative, and f positive on R.
15. f positive, f negative, and f positive on [0, 3].
16. f negative, f negative, and f negative on [0, 3].
19. f negative, f negative, and f negative on R.
20. f negative, f positive, and f positive on R.
Skills
In Exercises 21–28, graphs of f , f , or f are given. Whichever
is shown, sketch graphs of the remaining two functions.
Label the locations of any roots, extrema, and inflection points
on each graph.
Graph of f
22.
Graph of f
21.
8
y
10
4
y
5
2
2
4
x
6
3
4
2
1
x
1
5
8
10
23.
4
y
y
8
2
4
2
x
2
2
4
4
3 2 1
6
2
3
x
Graph of f y
y
32. f (x) =
33. f (x) =
1
3 − 2e x
34. f (x) = e x (x 2 − x − 1)
36. f (x) = cos(π x)
37. f (x) = arctan x
38. f (x) = sin−1 x 2
39. f (x) = sin(cos−1 x)
40. f (x) = cos(sin−1 x)
41. f (x) = (x − 2)4
42. f (x) = (x − 3)3 (x − 1)
43. f (x) = x 4 − 2x 3 − 5
1
1 + x2
√
x
46. f (x) =
x−2
x
48. f (x) =
ln x
2
51. f (x) = sin x −
9
6
3
5 4 3 2 1
27.
1
Graph of f
3 2 1
3
x
2
1
2
3
28.
52. f (x) = 3 cos
f’
2
2
4
2x
1 − 2x
f ’’
2
1
3
π
4
y
3
2
50. f (x) =
1
2
3
4
5
x
1
1
2
3
4
5
x
54.
1
2
Use the second-derivative test to determine the local extrema
of each function f in Exercises 29–40. If the second-derivative
test fails, you may use the first-derivative test. Then verify your
2
0.5
4
6
π
x +5
2
For each set of sign charts in Exercises 53–62, sketch a possible
graph of f .
53.
Graph of f
y
2 1
1
x
6
44. f (x) =
1
x2 + x + 1
12
5
(x − 1)2
x+2
35. f (x) = cos(π(x + 1))
49. f (x) = e 1+2x−x
15
10
1 + x + x2
x2 + x − 2
47. f (x) = e 3x (1 − e x )
Graph of f 26.
30. f (x) = x 2 (x − 1)(x + 1)
31. f (x) =
45. f (x) =
4
8
25.
1
29. f (x) = (x − 2)2 (1 + x)
Use a sign chart for f to determine the intervals on which
each function f in Exercises 41–52 is concave up or concave
down, and identify the locations of any inflection points. Then
verify your algebraic answers with graphs from a calculator or
graphing utility.
Graph of f 24.
Graph of f
algebraic answers with graphs from a calculator or graphing
utility. (Note: These are the same functions that you examined with
the first-derivative test in Exercises 39–50 of Section 3.2.)
55.
2
2
CONFIRMING PAGES
f’
f ’’
f’
f ’’
20:18
TKmaster2010
WHF00153/FREE087-Taalman
276
56.
Chapter 3
November 24, 2012
Applications of the Derivative
2
2
1
2
f’
f ’’
62.
57.
DNE 0
1
2
DNE
58.
1
1
2
DNE
1
59.
3
2
0
1
60.
0
1
61.
3
1
4
2
1
1
1
1
f ’’
67. f (x) = x 3 − 2x 2 + x
68. f (x) = (1 − x)4 − 2
x
70. f (x) = 2
x +1
1
72. f (x) =
(x − 1)2 (x − 2)
f’
f ’’
f
f ’’
2
DNE
66. f (x) =
f’
3
f’
65. f (x) = x 3 + 3x 2
f ’’
1
DNE
64. f (x) = (x − 2)(x + 2)
f’
1
63. f (x) = x 2 + 3x
f
0
0
f
f ’’
69. f (x) = x 3 (x + 2)
71. f (x) =
√
x (4 − x)
x2 − 1
− 5x + 4
1−x
75. f (x) =
ex
x
e
77. f (x) =
x
73. f (x) =
1
DNE
DNE 1
2
f’
f
1
1
Sketch careful, labeled graphs of each function f in Exercises 63–82 by hand, without consulting a calculator or graphing utility. As part of your work, make sign charts for the signs,
roots, and undefined points of f , f , and f , and examine any
relevant limits so that you can describe all key points and
behaviors of f .
f ’’
DNE 0
1
f’
1
DNE 2
x2
74. f (x) =
1
x2 + 3
x2
x2 − x
− 3x + 2
76. f (x) = ln(x 2 + 1)
78. f (x) = e 3x − e 2x
79. f (x) = x 2/3 − x1/3
80. f (x) = cos 3 x −
81. f (x) = ( ln x)2 + 1
82. f (x) = sin(tan−1 x)
π
2
In Exercises 83–86, use the given derivative f to find any
local extrema and inflection points of f and sketch a possible
graph without first finding an formula for f .
83. f (x) = x 3 −3x 2 +3x
1
85. f (x) =
x
84. f (x) = x 4 − 1
86. f (x) = e x (x + 4)
Applications
87. Jason’s distance in miles north from the corner of Main
Street and High Street t minutes after noon on Tuesday is
given by the following function s(t):
Distance north of Main and High
y
1
20
1
2
40
60
80
x
(a) Find an interval on which Jason’s velocity is positive
and decreasing. Describe what Jason is doing over
this time interval.
(b) Find a time interval on which Jason is moving north
and his velocity is increasing. Describe what Jason is
doing over this time interval.
(c) Find a time interval on which Jason’s acceleration
and velocity are both negative. Describe what Jason
is doing over this time interval.
(d) At which time is Jason’s velocity at a minimum? What
is he doing at that moment?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.3
88. Suppose Juri drives for two hours and that his distance
from home in miles is given by the function s(t) shown in
the following figure.
Distance from home
y
100
75
The Second Derivative and Curve Sketching
277
where both x and h(x) are measured in miles. Find the
areas of the glacier where it is concave down and hence
where Ian will need to move away from the center to avoid
the crevasses.
90. Ian is a bit worried about taking a fall into a crevasse while
carrying a heavy pack and towing a heavy sled. He does
some tests on an old rope, dropping from a tree in his
backyard as shown below.
50
25
1
2
t
(a) Find a time interval on which Juri’s acceleration is positive. Is his velocity positive or negative on this interval?
Describe what Juri is doing over this time interval.
(b) The graph y = s(t) of Juri’s position has an inflection
point at t = 1 hour. What does this inflection point
say about Juri’s velocity at t = 1? About his acceleration at t = 1?
89. For Exercises 89 and 90, suppose that Ian is a climber
who is planning a path over a glacier in Canada’s Icefield
Range. Glaciers are a little like gelatin: They tend to form
cracks (crevasses) when their surfaces are concave down.
Cracks close up and travel is easy when they are concave
up, as shown in the figure.
h(x)
Using a map, Ian approximates the elevation of the glacier
on a line that runs up through its center as
h(x) = 1.2 + .0095x + 0.037x 2 − 0.0072x 3 + 0.00046x 4 ,
x(t)
t 1.2
He cannot measure the force on the rope, but he takes a
video, from which he can find his position (in feet below
the tree limb) at time t > 1.2 seconds as
x(t) = 20 + 7e −0.25t sin(4.7t − 5.8).
(a) Weight is a force, given by mass times acceleration.
Ian weighs 160 pounds, and the acceleration due to
gravity that causes his weight is 32 feet per square
second. What is Ian’s mass? (The units are called
“slugs.”)
(b) Recalling that acceleration is the second derivative
of position x(t), what is the force on Ian at any time
t > 1.2?
(c) Use a graphing calculator or other graphing utility to
make a graph of Ian’s acceleration over time. When is
the upward force on Ian the greatest? Note that since
we are measuring distance below the tree limb, in this
situation an upward force is negative. What is that
force?
Proofs
91. Prove part (b) of Theorem 3.10: If both f and f are differentiable on an interval I, and f is negative on I, then f is
concave down on I.
92. Prove Theorem 3.11 (b): If x = c is a critical point of f ,
both f and f are differentiable near x = c, and if f (c) is
negative, then f has a local maximum at x = c.
93. Prove that every quadratic function is either always
concave up or always concave down.
94. Prove that every cubic function (i.e., every function of the
form f (x) = ax 3 + bx 2 + cx + d for some constants a,
b, c, and d) has exactly one inflection point. (Note: It is
not enough just to show that the second derivative of any
cubic function has exactly one zero; you must also show that
the sign of the second derivative changes.)
95. Prove that if f is zero on an interval, then f is linear on
that interval.
Thinking Forward
Global extrema on an interval: The first-derivative test
can be used to show that the function f (x) = x 2 + 3x
3
2
Global extrema and derivatives: The first-derivative
test can be used to show that the function
1
and
3
has a local minimum at x = − . Is this a global min-
f (x) = x 3 − x 2 + x has a local maximum at x =
imum of the function? Is there a global maximum?
What are the global extrema (if any) if we consider the
function restricted to the interval [−3, 3]?
a local minimum at x = 1. Are either of these local
extrema also global extrema? Can the first or second
derivative tell you whether or not a local extremum is
a global extremum?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
278
3.4
Chapter 3
November 24, 2012
Applications of the Derivative
OPTIMIZATION
Comparing local extrema and endpoint behavior to find global extrema
Strategies for translating real–world problems into mathematical problems
Optimizing real–world quantities by solving global extrema problems
Finding Global Extrema
Many real-world problems involve optimization, that is, finding the global maximum and
minimum values of a function on an interval. For example, a company might be interested in the maximum amount of profit it can make or in the minimum cost of producing an object. In this section we will learn how to use limits, derivatives, and values to
find the global extrema of a function on an interval, so that we can solve such real-world
problems.
Recall from Section 0.4 that x = c is the location of a global maximum of a function f
on an interval I if f (c) ≥ f (x) for all x ∈ I. Sometimes local extrema are also global extrema,
and sometimes they are not. For example, the following four functions each have a local
maximum at the critical point x = 2:
f has a global maximum
on [1, 5] at x = 2
h has no global
maximum on [1, 5)
g has a global maximum
on [1, 5] at x = 5
y
y
k has no global
maximum on [1, 5)
y
y
4
4
4
4
2
2
2
2
1
2
5
x
1
2
5
x
1
2
5
x
1
2
5
x
Only in the first graph is the local maximum also a global maximum. In the graph of g
the global maximum on [1, 5] is instead at the right endpoint x = 5. In the last two graphs
there is no global maximum on the interval, because there is no point on either graph that
is higher than all other points on its graph.
Although derivatives can help us locate local extrema, they do not always provide
enough information to detect global extrema. The previous figures suggest that to find
a global extremum of a continuous function we must compare the following:
the values f (c) for each interior local extremum c ∈ I;
the values f (c) at any closed endpoints x = c of I;
the limits lim f (x) at any open endpoints or non-domain points x = c of I.
x→c
Whichever of these three values is largest determines the location, if any, of the global
maximum of f on I. Whichever is smallest determines the location, if any, of the global
minimum. For example, in the graph of f that we just looked at, the value at the local
extremum is f (2) = 2 and at the left and right endpoints we have f (1) = 0 and
f (5) = 0. Since the value is highest at x = 2, that point is the global maximum of f on
[1, 5]. A different thing happens with the graph of h; we again have h(2) = 2 at the
local extremum and h(1) = 0 at the left endpoint, but at the open right endpoint we have
lim− h(x) = 4. Since the limit at the open right endpoint is larger than the value at the
x→5
left endpoint and the values at any local extrema, the function h has no global maximum
on [1, 5).
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
Optimization
279
Translating Word Problems into Mathematical Problems
Finding the global extrema of a function f on an interval I is a mathematical problem that
we can tackle with function values, limits, and derivatives. Real–world problems are less
straightforward, since they are expressed in full sentences in which the variables, constants,
functions, and relationships are described in words instead of in mathematical notation.
To solve a real–world optimization problem we must first translate it into a mathematical
global extrema problem. Once we solve the mathematical problem, we can translate the
solution back into the real–world context. This translating procedure is illustrated in the
following diagram:
Word problem
in English
???
Answer
in English
Translate
Mathematical
problem
Translate
Solve
Mathematical
answer
Many students have difficulty with word problems because of the translation step, not
because of the mathematical solving step. As a simple example, suppose we wanted to
solve the following real–world problem:
Calvin throws a baseball straight into the air at 50 feet per second, releasing the
ball when his hand is 5 feet above the ground. How high did Calvin throw the
baseball?
To figure out how to answer this problem with calculus, we need to get some equations
and variables to work with. A good place to start is to list what you know and set variable
names as you go along:
Let s(t) be the height/position of the baseball at time t, in feet.
We know that the position equation will be of the form s(t) = −16t 2 + v0 t + s0 .
v0 = 50 feet per second is the initial velocity of the baseball.
s0 = 5 feet is the initial position of the baseball.
We seek the greatest height of the baseball during the time that it is in the air.
Solving s(t) = 0, we see that the ball hits the ground at t ≈ 3.22 seconds.
At some point while developing this list of facts it is good to collect the information into a
figure, such as this one:
?
s(t) 16t 2 v0 t s0
v0 50 ft/s
t0
s0 5 ft
s(t) 0
t ≈ 3.22
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
280
Chapter 3
November 24, 2012
Applications of the Derivative
Putting this all together, we see that the mathematical problem that we wish to solve is
Find the global maximum value of s(t) = −16t 2 + 50t + 5 on the interval [0, 3.22].
Compare this mathematical statement with our earlier real–world statement. After the hard
work of translating, we now have a fairly straightforward calculus problem to solve. Here
is an outline of the steps we could use to solve this optimization problem:
Take the derivative: s (t) = −32t + 50.
Find critical points: s (t) = 0 when t =
Test critical points:
x = 1.5625.
s (1.5625)
50
32
= 1.5625.
= −32 < 0, so s(t) has a local maximum at
Check height at local maximum: s(1.5625) ≈ 44.06.
Check heights at endpoint values: s(0) = 0 and s(3.22) ≈ 0.
Conclusion: The maximum value of s(t) on [0, 3.22] is approximately 44.06.
Translating back into the real–world context, we can say that Calvin threw the baseball to
a height of just over 44 feet.
In general, when translating a word problem we must identify any variables or functions and express them in mathematical notation with letters and symbols. With those
same letters and symbols we can construct labeled diagrams, formulas, and relationships in
mathematical notation. We must keep translating until we form a well-posed mathematical
problem that we know how to solve with the tools and techniques of calculus. In the context of optimization problems, this means that we must identify a specific function and a
specific interval on which we wish to find a global maximum or minimum value. At this
point we only know how to find global extrema of one-variable functions, so we will sometimes have to use constraint equations to reduce multivariable functions to single-variable
functions; see Example 2.
Examples and Explorations
EXAMPLE 1
Finding global extrema of a function on an interval
Find the global maximum and minimum values, if any, of the function f (x) = 2x 3 −
15x 2 + 24x + 20 on the interval (0, 6).
SOLUTION
Our goal will be to use the derivative to identify local extrema on the interior of the interval
and then compare those extrema with values or limits at the ends of the interval. In this
case the interval is open, so we will compare with limits at x → 0 and x → 6. Step by step,
we have
Take the derivative: f (x) = 6x 2 − 30x + 24 = 6(x 2 − 5x + 4) = 6(x − 1)(x − 4).
Find critical points: f (x) = 0 at x = 1 and x = 4.
Test critical points: f (x) = 12x − 30, so f (1) = −18 < 0 and f (4) = 18 > 0; thus
f has a local maximum at x = 1 and a local minimum at x = 4.
Check height at local extrema: f (1) = 31 and f (4) = 4.
Check limits at open endpoints:
lim f (x) = lim+ (2x 3 − 15x 2 + 24x + 20) = 20,
x→0+
x→0
lim− f (x) = lim− (2x 3 − 15x 2 + 24x + 20) = 56.
x→6
x→6
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
Optimization
281
The local maximum value of f (1) = 31 is exceeded by the limit of f at the right endpoint,
so the function f has no global maximum on (0, 6). The local minimum value of f (4) = 4 is
less than the limit at either end of the interval and therefore is also a global minimum of f
on (0, 6). The graph of f on (0, 6) looks like this:
Global minimum at x = 4,
no global maximum on (0, 6)
y
56
31
20
4
1
EXAMPLE 2
4
6
x
Optimizing a function on an interval given a constraint
Farmer Joe wants to build a rectangular chicken pen for his chickens. He wants to build
the pen so that it has the largest area possible, and he has only 100 feet of chicken-wire
fencing. What dimensions should he use for the pen?
SOLUTION
Let l and w represent the length and width of the rectangular pen, in feet. We know that
the perimeter P of the pen must be 100 feet, since that is how much chicken-wire fencing
Farmer Joe has; therefore P = 2w + 2l = 100. We are interested in maximizing the area
A = lw of the pen. The following diagram summarizes this information:
Constraint: P = 2w + 2l = 100
w
Maximize: A = l w
l 50 w
At this point the function A that we wish to maximize is written in terms of two variables, but our calculus techniques work only for functions of one variable. Luckily, we can
use the constraint P = 2w + 2l = 100 to solve for one variable in terms of the other; this
will reduce A to a one-variable function that we can optimize. Solving for l in terms of w,
we have
P = 2w + 2l = 100
=⇒
2l = 100 − 2w
=⇒
l = 50 − w.
Using this equation, we can write the area function entirely in terms of w:
A = lw = (50 − w)w = 50w − w2 .
We now have a one-variable function A(w) = 50w − w2 to maximize, but on what interval? We need to know which values of w we are willing to consider in this problem.
Clearly we must have w ≥ 0, since the width of the chicken pen cannot be negative. What
is the upper bound on w? The key is to realize that the width w is the largest when the
length l is the smallest. The smallest value that l could have is l = 0, and when l = 0, we
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
282
Chapter 3
November 24, 2012
Applications of the Derivative
have 0 = 50 − w, or w = 50. Therefore we must have w ≤ 50. We have now determined
that the underlying mathematical problem for this word problem is
Find the global maximum of A(w) = 50w − w2 on the interval [0, 50].
Now we just have a simple mathematical problem to solve, and we follow the usual
procedure for finding a global extremum on an interval:
Take the derivative: A (w) = 50 − 2w.
Find critical points: A (w) = 0 at w = 25.
Test critical points: A (w) = −2, so A (25) = −2, and by the second-derivative test,
A(w) has a local maximum at w = 25.
Check height at local extremum: A(25) = 625.
Check heights at closed endpoints: A(0) = 0 and A(50) = 0.
From this work we see that w = 25 is indeed the global maximum of A(w) on the interval
[0, 50]. Thus the width of the pen should be w = 25 feet long. Since l = 50 − w, the length
of the pen should also be l = 25 feet long, and the pen is square. The final answer to the
word problem is therefore that Farmer Joe should build a square chicken pen where each
side is 25 feet long.
EXAMPLE 3
Modeling and minimizing a cost function
You work for a company that makes jewelry boxes. Your boss tells you that each jewelry box
must have a square base and an open top and that you can spend $3.75 on the materials
for each box. The people in production tell you that the material for the sides of the box
costs 2 cents per square inch while the reinforced material for the base of the box costs
5 cents per square inch. What is the largest volume jewelry box that you can make and still
stay within budget?
SOLUTION
From a quick reading of the problem we can see that we wish to maximize the volume of
a jewelry box given certain monetary constraints. Let’s begin with a simple picture, assign
variable names, and list what we know in mathematical language. Suppose x is the length
of the sides of the square base and y is the height of the jewelry box, both measured in
inches. The material to make each of the four sides of the box will cost 0.02(xy) dollars,
and the material to make the base of the box will cost 0.05(x 2 ) dollars. Therefore we must
have 0.02(4xy) + 0.05(x 2 ) = 3.75, since we have $3.75 to spend on materials. We wish to
maximize the volume V = x 2 y of the box:
x
0.02(xy) dollars
y
0.05(x 2 ) dollars
y
x
x
x
x
Constraint: 0.02(4xy) + 0.05(x 2 ) = 3.75
Maximize: V = x 2 y
Again, the function to be optimized is written in terms of two variables at this point,
but we can use the constraint to solve for one variable in terms of the other. It is easiest to
solve for y in terms of x:
0.02(4xy) + 0.05(x 2 ) = 3.75 =⇒ 0.08xy = 3.75 − 0.05x 2 =⇒ y =
CONFIRMING PAGES
3.75 − 0.05x 2
.
0.08x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
283
Optimization
We can use this equation for y to write the volume function V in terms of one variable.
Doing this and then simplifying as much as possible, we have
V = x 2y = x 2
3.75 − 0.05x 2
0.08x
=
1
x(3.75 − 0.05x 2 ) = 46.875x − 0.625x 3 .
0.08
Now we must determine the appropriate domain for V(x) = 46.875x − 0.625x 3 in the
context of this problem. Clearly the smallest that the length x can be is zero. To find
the largest that the base side length x can be, consider that the smallest possible value
of the height of the box is y = 0. When y = 0, we have
0=
3.75 − 0.05x 2
0.08x
=⇒ 0 = 3.75 − 0.05x 2 =⇒ x =
√
3.75
= 75.
0.05
Note that we do not consider the negative square root of 75, since we know that the length
x must be nonnegative. We have now completely translated the original word problem into
the following mathematical problem:
Find the global maximum value of V(x) = 46.875x − 0.625x 3 on the interval [0,
√
75 ].
To solve this problem we will find all the local interior extrema of V(x) in the interval
√
[0, 75 ] and compare their values with the values of V at the endpoints of the interval. The
steps are the same as those from previous examples:
Take the derivative: V (x) = 46.875 − 3(0.625)x 2 .
Find
√ critical points: V (x) = 0 when x = ±5, but only x = 5 is in the interval
[0,
75 ].
Test critical points: V (5) = −2(3)(0.625)(5) is negative, so by the second-derivative
test, V(x) has a local maximium at x = 5.
Check height at local extremum: V(5) = 156.25.
√
Check heights at closed endpoints: V(0) = 0 and V( 75 ) = 0.
We now see that x = 5 is not only the √
location of a local maximum, but in fact the location
of the global maximum of V(x) on [0, 75 ]. Therefore the largest jewelry box that we can
make with the given cost restrictions has volume 156.25 cubic inches.
CHECKING
THE ANSWER
When x = 5, we must have y =
3.75 − 0.05(5)2
0.08(5)
= 6.25. To algebraically check some of the
work we have done, we can verify that, with these values for x and y, the cost of producing
the jewelry box is indeed 0.02(4xy) + 0.05(x 2 ) = $3.75. We can check our work regarding
the optimization of V and the choice of endpoints for the interval by graphing V(x) =
46.875x − 0.625x 3 , as shown next. From this graph, it does seem
√ reasonable that at x = 5
we have a global maximum of 156.25 and that the point x = 75 ≈ 8.66 should be the
right end of the interval for the problem.
160
9
0
0
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
284
Chapter 3
EXAMPLE 4
November 24, 2012
Applications of the Derivative
The fastest way to put out a flaming tent
While you are on a camping trip, your tent accidentally catches fire. Luckily, you happen
to be standing right at the edge of a stream with a bucket in your hand. The stream runs
east–west, and the tent is 40 feet north of the stream and 100 feet farther east than you are,
as shown here:
40 feet
100 feet
You can run only half as fast while carrying the full bucket as you can empty handed, and
thus any distance travelled with the full bucket is effectively twice as long. What is the
fastest way for you to get water to the tent?
SOLUTION
On the one hand, notice that if you were to get water immediately and run diagonally to
the tent, you would run the entire distance with a full bucket of water at half your normal
speed. On the other hand, if you were to run along the side of the stream until you were
directly south of the tent, then get water and run north to the tent, you would have a
lot of total distance to run. Clearly, it would be more efficient for you to run along the
side of the stream for a while, fill the bucket, and then run diagonally to the tent. The
question is, how far should you run along the side of the stream before you stop to fill up
the bucket?
We need some mathematical notation to get our heads around this problem. Suppose x
is the distance you will run along the stream before filling your bucket. The distance you will
run with the full bucket is then the hypotenuse of a right triangle withlegs of lengths 40 feet
and 100 − x feet. By the Pythagorean theorem, you will have to run 402 + (100 − x)2 feet
with a full bucket of water, as shown here:
40 2 (100 x)2
x
40 feet
100 x
100 feet
Since it takes you twice as long to run with a full bucket of water, the total effective distance
you will have to run is
D(x) = x + 2 402 + (100 − x)2 .
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
Optimization
285
This is the function we want to minimize. What is the interval of x-values that we are
interested in? From the diagram we can see that x must be between 0 and 100 feet, since
clearly you wouldn’t want to run away from the tent or past the tent. The endpoint cases
x = 0 and x = 100 correspond to the two special cases we discussed earlier: The case when
you get water immediately and the case when you run the full 100 feet before getting water.
We have now translated the word problem into the following mathematical optimization
problem:
Find the global minimum value of D(x) = x + 2 402 + (100 − x)2 on the interval [0, 100].
Once again this is a straightforward global extremum problem, and we follow the same
steps as previously; but this time the mathematics is a little more involved:
Take the derivative: By applying the chain rule twice, we see that the derivative of
D(x) is
D (x) = 1 + 2
1
(402 + (100 − x)2 )−1/2 (2(100 − x)(−1)).
2
Find critical points: We need to simplify D (x) before we attempt to find its zeroes
or the values where it does not exist. With a bit of algebra we can write D (x) in the
form of a quotient:
−2(100 − x)
402 + (100 − x)2 − 2(100 − x)
D (x) = 1 + =
.
402 + (100 − x)2
402 + (100 − x)2
From this quotient form it is easy to pick out any places where D (x) is zero or
undefined. The denominator of D (x) is never zero, and thus D (x) always exists. To
find the values of x for which D (x) = 0, we set the numerator equal to 0 and solve:
402 + (100 − x)2 − 2(100 − x) = 0
402 + (100 − x)2 = 2(100 − x)
402 + (100 − x)2 = 4(100 − x)2
2
1600 + 10, 000 − 200x + x = 40, 000 − 800x + 4x
← square both sides
2
0 = 3x 2 − 600x + 28400
40 √
3.
x = 100 ±
3
← quadratic formula
40 √
3 ≈ 123.1 is not in the interval [0, 100], but x =
Note that x = 100 +
3
40 √
3 ≈ 76.9 is. In addition, recall that squaring both sides of an equation
100 −
3
40 √
sometimes leads to extraneous, or false, solutions; interestingly, x = 100 +
3
3
happens to be one of those false solutions. Therefore the only critical point of D(x)
40 √
is x = 100 −
3 ≈ 76.9.
3
Test critical points: The first-derivative test can be applied to show that D(x) has a
local minimum at x = 100 −
40 √
3
3
∈ [0, 100].
40 √
Check height at local extremum: D(100 −
3 ) ≈ 169.282.
3
Check heights at closed endpoints: D(0) ≈ 215.407, D(100) = 180.
The work that we just did shows that the global minimum of D(x) on [0, 100] is at x =
40 √
100 −
3. This means that the quickest way to bring water to the burning tent is to run
3
40 √
for 100 −
3 ≈ 76.9 feet along the side of the stream, then fill the bucket with water and
3
run diagonally to the tent.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
286
Chapter 3
TEST YOUR
? UNDERSTANDING
November 24, 2012
Applications of the Derivative
Why can’t the derivative necessarily detect the global minimum of a function f on an
interval [a, b] if that minimum happens to occur at an endpoint of the interval?
Suppose a function f satisfies lim f (x) = −∞. Why does this mean that f has no global
x→3−
minimum on [0, 3]. What can you say about any global minima of f on [0, 6]?
Suppose a function f satisfies lim f (x) = ∞. What can you say about any global
x→1+
maxima of f on [−2, 4]?
What types of real–world problems translate into mathematical problems in which we
must find the global maximum or minimum of a function on an interval?
What are the general steps for solving an optimization word problem?
EXERCISES 3.4
Thinking Back
Local and global extrema: Use mathematical notation, including inequalities as used in the definition of local and global
extrema, to express each of the following statements.
On the interval [−3, 5], f has a local maximum at
x = 2.
On the interval [0, 2], f has a global maximum at
x = −1.
Critical points: Find the critical points of each of the following
functions.
3x − 2
x−1
π
x
f (x) = sin
2
f (x) =
f (x) =
1
√
1+ x
f (x) = e x (x − 2)
On the interval [−4, 4], f has a global minimum at
x = 0.
On the interval [0, 5], f has no global minimum.
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: Every local maximum is a global
maximum.
(b) True or False: Every global minimum is a local
minimum.
(c) True or False: If f has a global maximum at x = 2
on the interval (−∞, ∞), then the global maximum
of f on the interval [0, 4] must also be at x = 2.
(d) True or False: If f has a global maximum at x = 2 on
the interval [0, 4], then the global maximum of f on
the interval (−∞, ∞) must also be at x = 2.
(e) True or False: If f is continuous on an interval I, then
f has both a global maximum and a global minimum
on I.
(f) True or False: Suppose f has two local minima on the
interval [0, 10], one at x = 2 with a value of 4 and one
at x = 7 with a value of 1. Then the global minimum
of f on [0, 10] must be at x = 7.
(g) True or False: If f has no local maxima on (−∞, ∞),
then it will have no global maximum on the interval
[0, 5].
(h) True or False: If f (3) = 0, then f has either a local minimum or a local maximum at x = 3.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) The graph of a function with a local minimum at x = 2
but no global minimum on [0, 4].
(b) The graph of a function with no local or global
extrema on (−3, 3).
(c) The graph of a function whose global maximum on
[2, 6] does not occur at a critical point.
3. When you try to find the local extrema of a function f on
an interval I, one of the first steps is to find the critical
points of f . Explain why these critical points of f won’t
help you locate any “endpoint” extrema.
4. Explain why you can’t find the global maximum of a function f on an interval I just by finding all the local extrema
of f and then checking to see which one has the highest
value f (c).
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
5. Suppose f is a function that is defined and continuous on
an open interval I. Will the endpoints of I always be local
extrema of f ? Will f necessarily have a global maximum
or minimum in the interval I? Justify your answers.
6. Suppose f is a function that is defined and continuous on
a closed interval I. Will the endpoints of I always be local
extrema of f ? Will f necessarily have a global maximum
or minimum in the interval I? Justify your answers.
7. Suppose f is a function that is discontinuous somewhere
on an interval I. Explain why comparing the values of any
local extrema of f on I and the values or limits of f at the
endpoints of I is not in general sufficient to determine the
global extrema of f on I.
8. Given the following graph of f , graphically estimate the
global extrema of f on each of the six intervals listed:
y
6
5
4
3
2
1
4
(a) [0, 4]
(d) [0, ∞)
2
2
4
(b) [2, 5]
(e) (−∞, 0]
6
x
(c) (−2, 1)
(f) (−∞, ∞)
10. Given the following graph of f , graphically estimate the
global extrema of f on each of the six intervals listed:
y
(a) [−2, 4]
(d) (0, 4]
287
9. Given the following graph of f , graphically estimate the
global extrema of f on each of the six intervals listed:
y
40
8
30
6
20
4
10
2
3 2 1
10
Optimization
1
2
3
4
5
x
3 2 1
2
1
2
3
x
4
(b) (−2, 4)
(e) [0, 4)
(c) (−1, 1)
(f) (−∞, ∞)
(a) (−1, 1)
(d) [0, 2]
(b) [2, ∞)
(e) (1, ∞)
(c) [−2, −1)
(f) (−∞, ∞)
Skills
Find the locations and values of any global extrema of each
function f in Exercises 11–20 on each of the four given intervals. Do all work by hand by considering local extrema
and endpoint behavior. Afterwards, check your answers with
graphs.
11. f (x) = 2x 3 − 3x 2 − 12x, on the intervals
(a) [−3, 3]
(b) [0, 3]
(c) (−1, 2]
(b) [−2, 2]
3
(c) (−3, 1]
15. f (x) =
(b) [−2, 0]
(b) (0, 3)
(d) (−1, 3)
(c) [−1, 1]
(d) [0, 3]
(c) [1, 2]
(c) [−1, 1)
(d) [0, ∞)
, on the intervals
(b) [0, 10]
(c) [0, 3.5]
(a) [−2, 2]
(b) (0, 3)
(c) [0, ∞)
(d) (3, ∞)
(d) (−∞, 0]
(c) [1, 10]
Find dimensions for each shape in Exercises 21–24 so that the
total area enclosed is as large as possible, given that the total
edge length is 120 inches. The rounded shapes are half-circles,
and the triangles are equilateral.
21.
22.
23.
24.
(d) [0, ∞)
16. f (x) = √
+ 3, on the intervals
x2 + 1
(b) [−5, 2)
−1/2
(d) (−1, 1)
4
(a) (−5, 2]
(d) (0, 1)
20. f (x) = e x (x − 2), on the intervals
1
√ , on the intervals
1+ x
(a) [0, 3]
(b) (−2, 2)
19. f (x) = (x − 4x + 3)
(a) [0, 4]
(c) [0, 1)
π
x , on the intervals
2
4
(a) [−1, 1] (b) (−1, 1) (c) (−3, 0]
3x − 2
, on the intervals
14. f (x) =
x−1
(a) [0, 2]
(b) [0, 4)
(a) [−2, 2]
(d) (−2, 1)
13. f (x) = −12x + 6x + 4x − 3x , on the intervals
2
(a) [0, 4]
18. f (x) = sin
2
12. f (x) = 3x 4 + 4x 3 − 36x 2 , on the intervals
(a) [−5, 5]
17. f (x) = x 3/2 (3x − 5), on the intervals
(d) [0, 20]
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
288
Chapter 3
November 24, 2012
Applications of the Derivative
Use optimization techniques to answer the questions in
Exercises 25–30.
30. Find the volume of the largest cylinder that fits inside a
sphere of radius 10.
25. Find two real numbers x and y whose sum is 36 and
whose product is as large as possible.
26. Find two real numbers x and y whose sum is 36 and
whose product is as small as possible.
In Exercises 31–34, find the point on the graph of the function
f that is closest to the point (a, b) by minimizing the square of
the distance from the graph to the point.
27. Find real numbers a and b whose sum is 100 and for which
the sum of the squares of a and b is as small as possible.
28. Find the area of the largest rectangle that fits inside a
circle of radius 4.
31. f (x) = 3x + 1 and the point (−2, 1)
32. f (x) = x 2 and the point (0, 3)
33. f (x) = x 2 − 2x + 1 and the point (1, 2)
√
34. f (x) = x 2 + 1 and the point (2, 0)
29. Find the area of the largest rectangle that fits inside a
circle of radius 10.
Applications
A farmer wants to build four fenced enclosures on his farmland for his free-range ostriches. To keep costs down, he
is always interested in enclosing as much area as possible
with a given amount of fence. For the fencing projects in
Exercises 35–38, determine how to set up each ostrich pen so
that the maximum possible area is enclosed, and find this
maximum area.
35. A rectangular ostrich pen built with 350 feet of fencing
material.
36. A rectangular ostrich pen built along the side of a river (so
that only three sides of fence are needed), with 540 feet
of fencing material.
37. A rectangular ostrich pen built with 1000 feet of fencing
material, divided into three equal sections by two interior fences that run parallel to the exterior side fences, as
shown next at the left.
Ostrich pen with three sections
Ostrich pen with six sections
39. A rectangular habitat with a 20-foot-wide nesting hutch
along one side (so that fencing is not needed along those
20 feet), as shown next at the left.
Arena-style habitat
Rectangular habitat with hutch
nesting
hutch
20 feet
40. An arena-style habitat whose front area is a semicircle
and whose back area is rectangular, as shown previously
at the right.
41. A trapezoid-shaped habitat whose angled side is an enclosed walkway for zoo patrons (so that no fencing is
needed along the walkway), where the walkway makes
an angle of 60◦ with the right fence, as shown next at the
left.
Habitat along walkway
38. A rectangular ostrich pen that is divided into six equal sections by two interior fences that run parallel to the east
and west fences, and another interior fence running parallel to the north and south fences, as shown previously
at the right. The farmer has allotted 2400 feet of fencing
material for this important project.
You are in charge of constructing a zoo habitat for prairie
dogs, with the requirement that the habitat must enclose
2500 square feet of area and use as little border fencing
as possible. For each of the habitat designs described in
Exercises 39–42, find the amount of border fencing that the
project will require.
Habitat with mural
60
60
60
42. An arena-style trapezoid-shaped habitat whose long
back side is a wall with a landscape mural (so no fencing is
needed along the back wall), where the back wall makes
an angle of 60◦ with the slanted side fences, as shown
previously at the right.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.4
Alina wants to make keepsake boxes for her two best friends.
She doesn’t have a lot of money, so she wants to make each
box described in Exercises 43–44 so that it holds as much as
possible with a limited amount of material.
43. For Jen, Alina wants to make a box with a square base
whose sides and base are made of wood and whose top
is made of metal. The wood she wants to use costs 5 cents
per square inch, while the material for the metal top costs
12 cents per square inch. What is the largest possible box
that Alina can make for Jen if she only has $20.00 to spend
on materials?
Wood and metal box
Velvet-lined box
0.12 per sq. in.
Optimization
289
47. Linda also needs to mail some architectural plans, which
must be shipped in a cylindrical container. What is the
largest volume that her package can have? What is the
largest surface area that her package can have? (Hints:
The volume of a right circular cylinder with radius r and
height h is V = πr 2 h; the total surface area of such a cylinder
is SA = 2π rh + 2πr 2 .)
For Exercises 48–50, consider a toy car that moves back and
forth on a long, straight track for 4 minutes in such a way that
the function s(t) = 96t − 84t 2 + 28t 3 − 3t 4 describes how far
the car is to the right of the starting point, in centimeters, after
t minutes.
48. When is the toy car farthest away from the starting point?
Does it ever return to the starting point? For how long
does this model make sense?
49. Within the domain of your model, when is the toy car
moving fastest to the right? When is the toy car moving
fastest to the left?
50. When is the toy car accelerating the fastest to the right?
When is the toy car accelerating fastest to the left?
0.05 per sq. in.
l 2w
w
44. For Eliza, Alina wants to make a rectangular box whose
base is twice as long as it is wide. This box will be lined on
the entire inside with velvet and in addition the outside
of the top of the box is to be lined in velvet. If Alina has
240 square inches of velvet, how can she make Eliza’s box
so that it holds as many keepsakes as possible?
The U.S. Postal Service ships a package under large-package
rates if the sum of the length and the girth of the package is greater than 84 inches and less than or equal to 108
inches. The length of a package is considered to be the length
of its longest side, and the girth of the package is the distance around the package perpendicular to its length. In each
of Exercises 45–47, Linda wants to ship packages under the
USPS large-package rates.
45. Linda needs to mail a rectangular package with square
ends (in other words, with equal width and height). What
is the largest volume that her package can have? What is
the largest surface area that her package can have?
Suppose you have a 10-inch length of wire that you wish to
cut and form into shapes. In each of Exercises 51–53 you will
determine how to cut the wire to minimize or maximize the
area of the resulting shapes.
51. Suppose you wish to make one cut in the wire and use
the two pieces to form a square and a circle. Determine
how to cut the wire so that the combined area enclosed
by the square and the circle is (a) as small as possible and
(b) as large as possible.
52. Suppose you wish to make one cut in the wire and use
the two pieces to form a square and an equilateral triangle. Determine how to cut the wire so that the combined
area is (a) as small as possible and (b) as large as possible.
53. Suppose you wish to make one cut in the wire and use
the two pieces to form a circle and an equilateral triangle.
Determine how to cut the wire so that the combined area
of these two shapes is (a) as small as possible and (b) as
large as possible.
In each situation described in Exercises 54–62, set up and solve
a global extrema problem that solves the given real–world
optimization problem.
54. Alina needs to make a flyer for her band’s concert. The
flyer must contain 20 square inches of printed material
and for design purposes should have side margins of
1 inch and top and bottom margins of 2 inches. What size
paper should Alina use in order to use the least amount
of paper per flyer as possible?
h
w
l
46. Linda’s second package must be rectangular and 40
inches in length. What is the largest volume that her
package can have? What is the largest surface area that
her package can have?
55. Your company produces cylindrical metal oil drums that
must each hold 40 cubic feet of oil. How should the oil
drums be constructed so that they use as little metal as
possible? Can they be constructed to use as much metal
as possible?
56. An airplane leaves Chicago at noon and travels south at
500 miles per hour. Another airplane is travelling east towards Chicago at 650 miles per hour and arrives at 2:00
p.m.. When were these two airplanes closest to each other,
and how far apart were they at that time?
CONFIRMING PAGES
20:18
WHF00153/FREE087-Taalman
290
Chapter 3
November 24, 2012
Applications of the Derivative
arrives at
2:00 P.M.
Chicago
closest distance
departed at
12:00 P.M.
57. The cost of the material for the top and bottom of a cylindrical can is 5 cents per square inch. The material for the
rest of the can costs only 2 cents per square inch. If the can
must hold 400 cubic inches of liquid, what is the cheapest
way to make the can? What is the most expensive way?
58. Consider the can-making situation in the previous exercise, but suppose that the cans are made with open tops.
If each can must hold 400 cubic inches of liquid, what
is the cheapest way to make the cans? What is the most
expensive way?
pe
rf
t
59. A steam pipe must be buried underground to reach from
one corner of a rectangular parking lot to the diagonally
opposite corner. The dimensions of the parking lot are
500 feet by 800 feet. It costs 5 dollars per foot to lay steam
pipe under the pavement but only 3 dollars per foot to
lay the pipe along one of the long edges of the parking
lot. Because of nearby sidewalks, the pipe cannot be laid
along the 500-foot sides of the parking lot. How should
the steam pipe be buried so as to cost as little as possible?
(a) What price should you charge to sell the greatest
number of velvet Elvis paintings, and how many
could you sell at that price? For what price would
you sell the least number of paintings, and how many
would you sell?
(b) Write down a function that predicts the revenue R(c),
in dollars, that you will earn in a year if you charge
c dollars per painting. (Hint: Try some examples first;
for example, what would your yearly revenue be if you
charged $10.00 per painting? What about $50.00? Then
write down a function that works for all values of c.)
(c) What price should you charge to earn the most
money, and how much money would you earn? What
price per painting would cause you to make the least
amount of money in a year, and how much money
would you make in that case?
(d) Explain why you do not make the most money at the
same price per painting for which you sell the most
paintings.
62. While you are on a camping trip, your tent accidentally
catches fire. At the time, you and the tent are both 50
feet from a stream and you are 200 feet away from the
tent, as shown in the diagram. You have a bucket with
you, and need to run to the stream, fill the bucket, and
run to the tent as fast as possible. You can run only half
as fast while carrying the full bucket as you can empty
handed, and thus any distance travelled with the full
bucket is effectively twice as long. Complete parts (a)–(f)
to determine how you can put out the fire as quickly as
possible.
500 feet
$5
TKmaster2010
200 feet
$3 per ft
800 feet
50 feet
60. Suppose you want to make an open-topped box out of a
4×6 index card by cutting a square out of each corner and
then folding up the edges, as shown in the figure. How
large a square should you cut out of each corner in order
to maximize the volume of the resulting box?
x
x
4 inches
6 inches
61. Your family makes and sells velvet Elvis paintings. After
many years of research you have found a function that
predicts how many paintings you will sell in a year, based
on the price that you charge per painting. You always
charge between $5.00 and $55.00 per painting. Specifically, if you charge c dollars per painting, then you can
sell N(c) = 0.6c2 − 54c + 1230 paintings in a year.
(a) Let x represent the distance from the point on the
stream directly “below” you to the point on the
stream that you run to. Sketch the path that you
would follow to run from your starting position, to
the point x along the stream, to the tent.
(b) Let D(x) represent the effective distance (counting
twice any distance travelled while carrying a full
bucket) you have to run in order to collect water and
get to the tent. Write a formula for D(x).
(c) Determine the interval I of x-values on which D(x)
should be minimized. Explain in practical terms what
happens at the endpoints of this interval, and calculate the value of D(x) at these endpoints.
(d) Find D (x), and simplify as much as possible. Are
there any points (in the interval I) at which D (x) is
undefined?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
(e) It is difficult to find the zeroes of D (x) by hand. Use
a graphing utility to approximate any zeroes of D (x)
in the interval I, and test these zeroes by evaluating
D at each one.
Related Rates
291
(f) Use the preceding information to determine the
minimum value of D(x), and then use this value to
answer the original word problem.
Proofs
63. Prove that the rectangle with the largest possible area
given a fixed perimeter P is always a square.
64. Prove that the most efficient way to build a rectangular
fenced area along a river—so that only three sides of
fencing are needed—is to make the side parallel to the
river twice as long as the other sides. You may assume
that you have a fixed amount of fencing material.
Thinking Forward
Consider the graph of the function f shown next. Define A(x)
to be the area of the region between the graph of f and
the x-axis from 0 to x. We will count areas of regions above
the x-axis positively and areas of regions below the x-axis
negatively.
From the graph of f , estimate all global maxima and
minima of A(x), if any.
It turns out that the function f whose graph is shown
is given by the formula f (x) = 12x 3 − 96x 2 + 180x and
that the area function A is given by the formula A(x) =
3x 4 − 32x 3 + 90x 2 . What surprising relationship do f
and A have?
Show that your answers for the local and global extrema of A(x) are reasonable by using optimization
techniques on the area function A(x) = 3x 4 − 32x 3 +
90x 2 .
A(x) is area under this graph on [0, x]
y
100
80
60
40
20
20
40
3.5
1
2
3
4
5
x
Use the graph to approximate the values of A(0), A(1),
A(2), A(3), A(4), and A(5).
From the graph of f , estimate all local maxima and
minima of A(x).
RELATED RATES
Using implicit differentiation to obtain relationships between rates
Formulas for volume, surface area, and relationships between side lengths of triangles
Techniques for solving related-rates problems
Related Quantities Have Related Rates
If two quantities that change over time are related to each other, then their rates of change
over time will also be related to each other. For example, consider an expanding circle.
Clearly the radius r = r(t) of the circle and the area A = A(t) of the circle are related: If you
know one of these quantities at some time t, then you also know the other, via the formula
dr
A = π r 2 . As the circle expands over time, the rate at which its radius increases is related
to the rate
dA
dt
dt
at which its area increases. We can find an equation that relates these rates
by applying implicit differentiation to the formula that relates the quantities r and A :
A = πr 2
d
d
(A(t)) = (π (r(t))2 )
dt
dt
dA
dr
= π 2r
dt
dt
dA
dr
= 2πr .
dt
dt
← relationship between A and r
← differentiate both sides with respect to t
← chain rule
← relationship between
dA
dr
and
dt
dt
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
292
Chapter 3
November 24, 2012
Applications of the Derivative
Notice that in this calculation we use A interchangeably with A(t) and r interchangeably
with r(t), employing the explicit function notation only when we need to be reminded that
A and r are functions of t.
According to the formula we just found, the rate
radius r at time t and the rate
dr
dt
dA
dt
depends on both the size of the
at which the radius is increasing. For example, suppose
that our expanding circle has an initial radius of r(0) = 2 inches at time t = 0, and that the
dr
radius then increases at a constant rate of = 3 inches per second. Then the formula we
dt
found tells us that at the instant the circle has a radius of 4 inches, the area of the circle is
increasing at a rate of
dA = 2π(4)(3) = 24π square inches per second.
dt r=4
In contrast, when the radius of the circle is 20 inches, the area of the circle is increasing at
the faster rate of
dA = 2π(20)(3) = 120π square inches per second.
dt r=20
Notice that although the rate of change of the radius of the circle is constant, the rate at
which the area is changing is not: As the circle gets larger and larger, changes to the radius
create larger and larger changes to the area. We could also determine the rate of change of
the area at a particular time, say, after t = 3 seconds. Since r(t) = 2 + 3t in our example,
at time t = 3 seconds, we have a radius of r(3) = 2 + 3(3) = 11 inches. The rate of change
of the area at this moment is
dA dA = 2π(11)(3) = 66π square inches per second.
=
dt t=3
dt r=11
Most related-rates problems involve two rates, one of which is known and one of which
you are asked to find. These rates will be related in some way that is determined by the way
the corresponding quantities are related. Translating a related-rates problem should result
in an equation that relates the two quantities whose rates you are interested in. You can
then implicitly differentiate the equation relating the quantities to get an equation relating
the rates, as we did at the outset of this section.
Volumes and Surface Areas of Geometric Objects
Many related-rates problems involve geometric quantities such as volume, area, and surface area. Many of the formulas for these quantities should already be familiar to you. For
example, here are some formulas for two-dimensional objects: A circle with radius r has
area A = πr 2 and circumference C = 2πr, a rectangle with length l and width w has area
A = l w and perimeter P = 2l + 2w, and any triangle with base b and height h has area
1
A = b h.
2
We now gather some three-dimensional geometric formulas for reference, although we
will not have the tools to prove them until much later in the book.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
THEOREM 3.12
293
Related Rates
Volume and Surface Area Formulas
The formulas that follow describe the volume V and surface area S of a rectangular box,
sphere, right circular cylinder, and right circular cone. The lateral (side) surface area L is
also given for the cylinder and cone.
(a) The volume and surface area of a rectangular box
of length x, width y, and height z are
V = xyz
z
S = 2xy + 2yz + 2xz
y
x
(b) The volume and surface area of a sphere of radius r
are
V=
4
πr 3
3
r
S = 4π r 2
(c) The volume, surface area, and lateral surface area of
a right circular cylinder of radius r and height h are
V = π r 2h
r
h
S = 2π r h + 2πr 2
L = 2π r h
(d) The volume, surface area, and lateral surface area of
a right circular cone of radius r and height h are
V=
1
πr 2 h
3
S = πr r 2 + h2 + π r 2
L = π r r 2 + h2
r
h
As always, whenever possible you should think rather than memorize. For example,
let’s walk through why the volume and surface area formulas for the cylinder make sense.
The volume of a cylinder is the product of the area πr 2 of its top and its height h. The
surface area of a cylinder is the sum of two things: the area of its curvy side and the areas of
its top and bottom circles. Notice that if we unrolled the curvy side, it would be a rectangle
whose width is the circumference 2π r of the top and bottom circles and whose height is
h. Therefore the lateral, or side, surface area of the cylinder is the product 2π r h. Of course,
the area of the top and bottom circles each have area πr 2 , so the top and bottom together
have area 2π r 2 . Thus the total surface area of the cylinder is 2π r h + 2πr 2 .
Similar Triangles
It is also common for related-rates word problems to involve the following two well-known
theorems concerning right triangles, which we present here for reference, without proof:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
294
Chapter 3
THEOREM 3.13
November 24, 2012
Applications of the Derivative
Two Theorems About Right Triangles
The following two theorems describe well-known relationships between the side
lengths of right triangles:
(a) The Pythagorean theorem states that if a right triangle
has legs of lengths a and b and hypotenuse of length c,
then:
a2 + b2 = c 2
(b) The law of similar triangles states that if two
right triangles have the same three angle measures,
so that one is just a scaled-up version of the other, then
the ratios of side lengths on one triangle are equal to
the ratios of corresponding side lengths on the other.
Specifically, with the side lengths shown in the diagram at the right, we have
h
H
= ,
b
B
d
D
= ,
b
B
c
b
a
D
d
b
h
H
B
d
D
=
h
H
The reason these theorems about triangles arise in related-rates problems is that both theorems give us ways to relate quantities that might change together over time. Finding an
equation that relates two quantities is often the first step in finding an equation that relates
the rates of change of those quantities.
Examples and Explorations
EXAMPLE 1
Relating quantities and rates
In each part that follows, write down an equation that relates the two given quantities.
Then use implicit differentiation to obtain a relationship between the rates at which the
following quantities change over time:
(a) the circumference C and the area A of a circle;
(b) the surface area S and the radius r of a cylinder with a fixed height of 4 units;
(c) the lengths a and b of the legs of a right triangle with hypotenuse of fixed length
7 units.
SOLUTION
(a) We know that a circle of radius r has circumference C = 2πr and area A = π r 2 . We
C
need to find an equation relating C and A. Since C = 2πr, we have r = ; substituting
2π
the right-hand side into A = πr 2 gives
A=π
C 2
1 2
=
C .
2π
4π
Now suppose that the circle is expanding or contracting, so that its area A = A(t)
and circumference C = C(t) are changing over time. By implicitly differentiating the
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
Related Rates
295
preceding equation with respect to t we have
dA
1
dC
1
dC
=
2C
=
C .
dt
4π
dt
2π
dt
(b) The formula for the surface area of a cylinder with radius r and height 4 is
S = 2π r(4) + 2π r 2 = 8π r + 2πr 2 .
By differentiating both sides we can express the relationship between the radius
r = r(t) and surface area S = S(t) if the cylinder changes size over time:
d(S)
dr
dr
dr
= 8π + 2π 2r = (8π + 4π r) .
dt
dt
dt
dt
(c) By the Pythagorean theorem, if a right triangle has legs of length a and b and hypotenuse of length 7, then
a2 + b2 = 7 2 .
If the triangle is changing shape or size over time in such a way that the hypotenuse
remains 7 units in length, then by implicit differentiation the rates of change of the leg
lengths a = a(t) and b = b(t) are related as follows:
2a
EXAMPLE 2
da
db
+ 2b
= 0.
dt
dt
Relating the changing radius and area of an expanding circle
Suppose a rock dropped into a pond causes a circular wavefront of ripples whose radius
increases at 3 inches per second. How fast is the area of the circle of ripples expanding at
the instant that the circle has a radius of 12 inches?
SOLUTION
Like many related-rates problems, this situation involves (1) two quantities that are related
and (2) known information about the rate of change of one of these quantities; we are then
asked to find information about the rate of change of the other quantity. In this case the
related quantities are the radius r = r(t) and area A = A(t) of the circle, which are related
by the formula A = πr2 . Both of these quantities change over time as the circle expands.
dr
We are given that the radius changes constantly at a rate of
= 3 inches per second and
asked to find the rate of change
summarized as follows:
dA
dt
dt
of the area at a particular moment. This information is
12 in.
t, seconds
r r(t)
3 in.
per sec
Relationship: A = πr 2
r(t) = radius
A(t) = area
area A A(t)
dr
=3
dt
dA Find:
dt r=12
Given:
At this point we are done translating the original real–world problem into a mathematical
related-rates problem that we know how to solve. To get a formula relating the rates of
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
296
Chapter 3
November 24, 2012
Applications of the Derivative
change of r(t) and A(t) we differentiate both sides of the area formula, yielding
By evaluating this equation for
dA
dt
dA
dt
dr
dt
= 2πr .
at r = 12 and using the given information that
dr
dt
= 3,
we see that at the moment that the circle has a 12-inch radius, it is expanding at a rate of
dA dr
= 2π (12) = 2π (12)(3) = 72 π square inches per second.
dt r=12
EXAMPLE 3
dt
Relating the changing volume and radius of an inflating balloon
Suppose a pink spherical party balloon is being inflated at a constant rate of 44 cubic inches
per second.
(a) How fast is the radius of the balloon increasing at the instant that the balloon has a
radius of 4 inches?
(b) How fast is the radius of the balloon increasing at the instant that the balloon contains
100 cubic inches of air?
SOLUTION
This is a related-rates problem because it involves two rates, namely, the rate at which the
balloon is being inflated and the rate of change of the radius of the balloon. We know
something about the first rate and wish to say something about the second. Suppose
r = r(t) is the radius of the balloon in inches after t seconds and V = V(t) is the volume
of the balloon in cubic inches after t seconds. The quantities r and V are related by the
4
volume equation V = πr 3 . We are given that the rate of change of the volume is constantly
dV
dt
3
= 44, and we want to find the rate of change
dr
dt
of the radius when r = 4 and
when V = 100. The following diagram summarizes this translation of the problem into
mathematical notation:
4 in.
t, seconds
r r(t)
Relationship: V =
r(t) = radius
V(t) = volume
4 3
πr
3
dV
= 44
dt
dr dr and Find: dt r=4
dt V=100
Given:
volume V V(t)
dV
To get an equation relating the rates
and
dt
tion that relates V and r with respect to t:
dr
dt
we differentiate both sides of the equa-
dV
4
dr
dr
= π 3r 2 = 4π r 2 .
dt
3
dt
dt
dV
is always equal to 44. Therefore we have the
We know from the given information that
dt
following formula concerning the rate of change of the radius of the balloon:
44 = 4πr 2
dr
.
dt
(a) To find the rate of change of the radius at the instant that the ballonhas a 4-inch radius,
dr we evaluate the preceding equation at r = 4 and then solve for :
dt r=4
dr 44
2 dr 44 = 4π (4 ) =⇒
≈ 0.2188 inch per second.
=
dt r=4
dt r=4
4π (16)
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
297
Related Rates
(b) For the second part we must find the rate at which the radius is changing at the moment that the volume of the balloon is 100 cubic inches. Our formula for the rate of
change of the radius of the balloon depends on r, not V, so we must first use the
4
equation V = πr 3 to determine the value of r when V = 100:
3
100 =
4
πr 3
3
r3 =
=⇒
100
(4/3) π
=⇒
r=
100
(4/3)π
1/3
≈ 2.879.
Using our previously developed formula, we find that when the volume of the balloon
is V = 100 cubic inches and thus the radius is approximately r ≈ 2.879 inches, the rate
of change of the radius of the balloon is
dr dr 44
44 = 4π(2.879)2 =⇒
=
≈ 0.422 inch per second.
2
dt V=100
EXAMPLE 4
dt V=100
4π (2.879)
The shadow of a person walking away from a streetlight
Matt is 6 feet tall and is walking away from a 10-foot streetlight at a rate of 3 feet per second.
As he walks away from the streetlight, his shadow gets longer. How fast is the length of
Matt’s shadow increasing when he is 8 feet from the streetlight?
SOLUTION
We are given the rate at which Matt walks away from the streetlight, and we wish to find
the rate of change of the length of Matt’s shadow. To find a relationship between these two
rates we will find a relationship between their underlying variables: the distance s between
Matt and the streetlight and the length l of Matt’s shadow. By the law of similar triangles,
s and l are related by the equation
10
s+l
6
l
= , as shown in the following diagram:
3 feet/sec
Relationship:
10 ft
s(t) = distance
6 ft
s s(t)
10
6
=
s+l
l
ds
=3
dt
dl Find: dt s=8
Given:
l(t) = length
l l(t)
ds
dl
To find the relationship between and we must implicitly differentiate the equation
dt
dt
relating s and l. We will simplify the equation first to make our lives easier:
10
6
=
s+l
l
=⇒
10l = 6(s + l )
=⇒
Differentiating both sides and then using the fact that
4
dl
ds
=6
dt
dt
=⇒
4
dl
= 6(3)
dt
ds
dt
4l = 6s.
= 3 gives us
=⇒
dl
= 4.5.
dt
Interestingly, we have just discovered that Matt’s shadow is increasing at a constant rate of
4.5 feet per second. In particular, when
Matt is 8 feet from the streetlight, the length of his
dl shadow is increasing at a rate of = 4.5 feet per second.
dt s=8
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
298
Chapter 3
November 24, 2012
Applications of the Derivative
d 3
(r )
dt
Suppose r = r(t) is a function of t. Why is
TEST YOUR
? UNDERSTANDING
not equal to 3r 2 ?
What are the formulas for the volumes of spheres, cylinders, and cones? What about
the formulas for surface area?
What do we mean when we say that two triangles are similar? What does the law of
similar triangles say?
In Example 2, why did we not label the diagram of the circle with the number 12? Why
did we use the variable r instead?
In Example 3 the radius of the balloon increases at a faster rate when the balloon is
smaller. Why does this make sense?
EXERCISES 3.5
Thinking Back
Using the chain rule: Given that r = r(t), s = s(t), and u = u(t)
are functions of t and that c and k are constants, find each of
the following derivatives.
d
(π u2 )
dt
d
(3r + 2s)
dt
d
(cu + rs)
dt
d
(k + cu3 )
dt
d
(cr 2 u)
dt
d
dt
c+s
k+u
Evaluation in Leibniz notation: Given that r = r(t), s = s(t), and
u = u(t) are functions of t, answer each of the following.
ds
ds If
= 3s2 − 4, find .
dt
dt
s=2
dr
dr If r 2 − 2r = 0, find .
dt
dt r=3
If 4 = 2u
du
du and u = 2 + 3t, find .
dt
dt t=4
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If a square grows larger, so that its side
length increases at a constant rate, then its area will
also increase at a constant rate.
(b) True or False: If a square grows larger, so that its side
length increases at a constant rate, then its perimeter
will also increase at a constant rate.
(c) True or False: If a circle grows larger, so that its radius increases at a constant rate, then its circumference will also increase at a constant rate.
(d) True or False: If a sphere grows larger, so that its radius
increases at a constant rate, then its volume will also
increase at a constant rate.
(e) True or False: The volume of a right circular cone is
one-third of the volume of the right circular cylinder
with the same radius and height.
(f) True or False: If V(r) is the volume of a sphere as a
function of its radius, and S(r) is the surface area of a
sphere as a function of its radius, then V (r) = S(r).
(g) True or False: If you unroll the side of a right
circular cylinder with radius r and height h, you get
a flat rectangle with height h and width 2πr.
(h) True or False: Given a right triangle with side lengths
a and c and hypotenuse of length b, we must have
a 2 + b2 = c 2 .
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Two triangles that are similar to the
√ right triangle with
legs of length 1 and hypotenuse 2.
(b) Two triangles with a hypotenuse of length 5.
(c) Two cylinders with a volume of 100 cubic units.
3. Give formulas for the volume and the surface area of a
cylinder with radius y and height s.
4. Give formulas for the volume and the surface area of a
cone with radius u and height w.
5. Give formulas for the volume and the surface area of a
cylinder whose radius r is half of its height h.
6. Give formulas for the volume and the surface area of a
cone whose height h is three times its radius r.
7. State the Pythagorean theorem and give an example of a
triangle that illustrates the theorem.
8. State the law of similar triangles and give an example of
a pair of triangles that illustrate this law.
9. If the volume and radius of a sphere are functions of time,
what is the relationship between the rate of change of the
volume of the sphere and the rate of change of the radius
of the sphere?
10. If the volume and radius of a cone are functions of time,
what is the relationship between the rate of change of the
volume of the cone and the rate of change of the radius
of the cone?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
11. If the volume and equator of a sphere are functions of
time, what is the relationship between the rate of change
of the volume of the sphere and the rate of change of the
equator of the sphere?
12. Suppose the side lengths x, y, and z of a rectangular box
are each functions of time.
(a) How is the rate of change of the volume of the box
related to the rates of change of x, y, and z?
(b) How is the rate of change of the surface area of the
box related to the rates of change of x, y, and z?
13. Suppose the radius r, volume V, and surface area S of a
sphere are functions of time t.
dV
dr
and
related?
dt
dt
dS
dr
and
related?
(b) How are
dt
dt
(a) How are
14. Suppose the radius r, volume V, and surface area S of
dV
dS
a sphere are functions of time t. How are
and
dt
dt
299
Related Rates
15. Suppose the radius r, height h, and volume V of a
dV
dr
cylinder are functions of time t. How is
related to
dt
dt
if the height of the cylinder is constant?
16. Suppose the radius r, height h, and volume V of a
dV
dh
cylinder are functions of time t. How is
related to
dt
dt
if the radius of the cylinder is constant?
17. Suppose the radius r, height h, and volume V of a
cylinder are functions of time t, and further suppose that
the height of the cylinder is always twice its radius. Write
dV
dh
in terms of h and .
dt
dt
18. Suppose the radius r, height h, and volume V of a cylinder are functions of time t, and further suppose that the
volume of the cylinder is always constant. Write
terms of r, h, and
dh
.
dt
dr
in
dt
related?
Skills
In Exercises 19–26, write down an equation that relates the
two quantities described. Then use implicit differentiation to
obtain a relationship between the rates at which the quantities
change over time.
19. The area A and perimeter P of a square.
26. The area A and hypotenuse c of a triangle that is similar
to a right triangle with legs of lengths 3 and 4 units and
hypotenuse of length 5 units.
Given that u = u(t), v = v(t), and w = w(t) are functions of
t and that k is a constant, calculate the derivative
df
of each
dt
20. The area A and perimeter P of an equilateral triangle.
function f (t) in Exercises 27–36. Your answers may involve u,
21. The surface area S and height h of a cylinder with a fixed
radius of 2 units.
22. The volume V and radius r of a cylinder with a fixed height
of 10 units.
v, w,
du dv dw
, ,
, k, and/or t.
dt dt dt
27. f (t) = u2 + kv
28. f (t) = u + v + w
29. f (t) = tv + kv
√
31. f (t) = 2v u + w
30. f (t) = kuvw
24. The volume V and height h of a cone with a fixed radius
of 3 units.
33. f (t) = w(u + t)2
34. f (t) =
25. The area A and hypotenuse c of an isosceles right triangle.
35. f (t) =
23. The surface area S and radius r of a cone with a fixed
height of 5 units.
ut + w
k
32. f (t) = 3u2 v + vt
w
uv
k
36. f (t) = 2
uw
Applications
37. A rock dropped into a pond causes a circular wave of ripples whose radius increases at 4 inches per second. How
fast is the area of the circle of ripples expanding at the instant that the radius of the circle is 12 inches? 24 inches?
100 inches? Explain why it makes sense that the rate of
change of the area increases as the radius increases.
38. A rock dropped into a pond causes a circular wave of
ripples whose radius increases at 6 inches per second.
How fast is the area of the circle of ripples expanding at
the instant that the area of the circle is 100 square inches?
200 square inches? 1000 square inches? Explain why it
makes sense that the rate of change of the area increases
as the area increases.
In Exercises 39–42, suppose the sides of a cube are expanding
at a rate of 2 inches per minute.
39. How fast is the volume of the cube changing at the
moment that the cube has a side length of 8 inches?
40. How fast is the volume of the cube changing at the
moment that the cube has a side length of 20 inches?
41. How fast is the volume of the cube changing at the
moment that the cube’s volume is 55 cubic inches?
42. How fast is the volume of the cube changing at the
moment that the area of the cube’s base is 10 square
inches?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
300
Chapter 3
November 24, 2012
Applications of the Derivative
In Exercises 43–45, consider a large helium balloon that is
being inflated at the rate of 120 cubic inches per second.
43. How fast is the radius of the balloon increasing at the
instant that the balloon has a radius of 12 inches?
44. How fast is the radius of the balloon increasing at the instant that the balloon contains 300 cubic inches of air?
45. How fast is the surface area of the balloon increasing at
the instant that the radius of the balloon is 15 inches?
In Exercises 46–48, suppose that Stuart is 6 feet tall and is
walking towards a 20-foot streetlight at a rate of 4 feet per
second. As he walks towards the streetlight, his shadow gets
shorter.
46. How fast is the length of Stuart’s shadow changing? Does
it depend on how far Stuart is from the streetlight?
47. How fast is the tip of Stuart’s shadow moving? Does it
depend on how far Stuart is from the streetlight?
48. How fast is the area of the triangle made up of Stuart’s
legs and his shadow changing? Is it increasing or decreasing as Stuart walks towards the streetlight?
53. How fast is the radius of the conical salt pile changing
when the height of the pile is 4 inches?
54. How fast is the height of the conical salt pile changing
when the radius of the pile is 2 inches?
55. How fast is the height of the conical salt pile changing
when the height of the pile is 4 inches?
56. Linda is still bored and is now pouring sugar onto the
floor. The poured-out sugar forms a conical pile whose
height is three-quarters of its radius and whose height
is growing at a rate of 1.5 inches per second. How fast
is Linda pouring the sugar at the instant that the pile of
sugar is 3 inches high?
57. Riley is holding an ice cream cone on a hot summer day.
As usual, the cone has a small hole at the bottom, and ice
cream is melting and dripping through the hole at a rate
of half a cubic inch per minute. The cone has a radius of
2 inches and a height of 5 inches. How fast is the height of
the ice cream changing when the height of the ice cream
in the cone is 3 inches?
2 in.
In Exercises 49–51, Alina props a 12-foot ladder against the
side of her house so that she can sneak into her upstairs bedroom window. Unfortunately, the ground is muddy because
of a recent rainstorm, and the base of the ladder slides away
from the house at a rate of half a foot per second.
r
5 in.
h
12 ft
4 feet
1
foot per second
2
49. How fast is the top of the ladder moving down the side of
the house when the base of the ladder is 4 feet from the
house?
50. How fast is the top of the ladder moving down the side of
the house when the base of the ladder is 10 feet from the
house?
51. How fast is the area of the triangle formed by the ladder,
the house, and the ground changing when the top of the
ladder is 6 feet from the ground?
In Exercises 52–55, Linda is bored and decides to pour an entire container of salt into a pile on the kitchen floor. She pours
3 cubic inches of salt per second into a conical pile whose
height is always two-thirds of its radius.
52. How fast is the radius of the conical salt pile changing
when the radius of the pile is 2 inches?
58. Suppose the width w of a rectangle is decreasing at a
rate of 3 inches per second while the height h of the rectangle is increasing at a rate of 3 inches per second. The
rectangle initially has a width of 100 inches and a height
of 75 inches.
(a) Find the rate of change of the area of the rectangle in
terms of its width and height.
(b) On what intervals do the variables w and h make
sense in this problem? On what time interval does
the problem make sense?
(c) When will the area of the rectangle be increasing, and
when will it be decreasing? Answer these questions
both in terms of the width and height of the rectangle
and in terms of time.
59. Suppose the length a of one leg of a right triangle is increasing at a rate of 4 inches per second while the length
b of its other leg is decreasing at a rate of 2 inches per
second. The triangle initially has legs of width a = 1 inch
and b = 10 inches.
(a) Find the rate of change of the area of the triangle over
time, in terms of its width and height.
(b) On what intervals do the variables a and b make
sense in this problem? On what time interval does
the problem make sense?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.5
(c) When will the area of the triangle be increasing, and
when will it be decreasing? Answer these questions
both in terms of the width and height of the triangle
and in terms of time.
60. Annie is paddling her kayak through the San Juan Islands
and is a quarter of a mile away from where she wants to
cross a channel. She sees a ferry in the channel approaching fast from her left, about 2 miles away. The ferry travels
at about 20 mph, while Annie can do about 3 mph if she
jams.
20 m.p.h.
301
(c) Use implicit differentiation to determine how fast the
distance between Annie and the ferry is decreasing
when she first sees the ferry.
(d) If Annie decides to jam across the channel, will the
ferry hit her?
61. The sun goes down at a rate of about 11 degrees per hour
in Colorado in the middle of summer. Ian finds himself contemplating this fact one evening while sitting at
Chasm Lake, below Long’s Peak in Colorado, watching
the sun descend behind the peak. The point on the ridge
where the sun is descending is at 13,200 feet. The lake is
at 11,710 feet. Ian is sitting 3,100 horizontal feet from the
ridge.
t, time
2 miles
ferry
Related Rates
13,200 feet
0.25 miles
3 kayak
m.p.h.
(a) To set up a model for this problem, suppose Annie
is travelling on the x-axis and is approaching the origin from the right. Suppose also that the ferry is travelling on the y-axis and is approaching the origin
from above. Then x = x(t) represents Annie’s position at time t and y = y(t) represents the ferry’s position. Given what you know about this problem, what
are x (t) and y (t)?
(b) Construct an equation in terms of x = x(t) and y =
y(t) that describes the distance between Annie and
the ferry at time t.
11 degrees
per hour
␪ ␪(t)
x x(t)
11,710 feet
3100 feet
(a) Suppose x = x(t) is the distance of the edge of the
shadow from the ridge at time t. This distance is related to the angle θ = θ (t) shown in the figure. Find
a formula for the speed x (t) of the shadow.
(b) Use the model from part (a) to determine how fast
the shadow of the ridge is moving when it reaches
Ian.
Proofs
62. Prove that the lateral surface area of a right circular cone
is equal to π rl, where r is the radius of the cone and
l is the length of the diagonal of the cone, that is, the
distance from the vertex of the cone to a point on its
circumference.
64. Prove that the rate of change of the volume of a cylinder
with fixed height with respect to its radius r is equal to the
lateral surface area of the cylinder. Why does it make geometric sense that the lateral surface area would be related
to this rate of change?
63. Prove that the rate of change of the volume of a sphere
with respect to its radius r is equal to the surface
area of the sphere. Why does it make geometric sense
that the surface area would be related to this rate of
change?
65. Suppose a right triangle has legs of lengths a and b and a
hypotenuse of length h and that this triangle is changing
size, so that the length of its hypotenuse does not change.
Prove that the ratio of the rates of change
b
a
da
db
and
is
dt
dt
equal to − .
Thinking Forward
Parametric curves: Imagine the curve traced in the xy-plane by
the coordinates (x, y) = (3z + 1, z2 − 4) as z varies, where the
parameter z is a function of time t.
Plot the points (x, y) in the plane that correspond to
z = −3, −2, −1, 0, 1, 2, and 3.
If the parameter z moves at 3 units per second and
z = 0 when t = 0, plot the points (x, y) in the plane
that correspond to t = 0, 1, 2, 3, and 4.
If the parameter z moves at 5 units per second,
find the instantaneous rate of change of the x- and
y-coordinates as the curve passes through the
point (7, 0).
If the x-coordinate moves at 5 units per second, find
the instantaneous rate of change of the y-coordinate
as the curve passes through the point (7, 0).
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
302
3.6
Chapter 3
December 11, 2012
Applications of the Derivative
L ’ H Ô P I T A L ’ S R U L E
0
∞
and
0
∞
L’Hôpital’s rule for calculating limits of the indeterminate forms
Rewriting limits of the indeterminate form 0 · ∞ so that L’Hôpital’s rule applies
Using logarithms to calculate limits of the indeterminate forms 00 , 1∞ , and ∞0
Geometrical Motivation for L’Hôpital’s Rule
0
As we have already seen, limits of the form are indeterminate. At first glance it is not
0
clear whether such a limit exists or what it might be equal to. In some cases we can resolve
the indeterminate form
0
0
with some algebra, such as in this example:
x2
x
0
= lim 2
= 2
= 0.
x→0 x 3 − x
x→0 x − 1
0 −1
lim
0
Other limits of the indeterminate form are not so easy to simplify. In particular, limits of
0
quotients that involve a mixture of different types of functions are usually more resistant to
algebra. For example, consider the limit
x2
.
x→0 2 x − 1
lim
As x → 0 we have x 2 → 0 and 2 x − 1 → 1 − 1 = 0, and thus this limit is of the indetermi0
nate form . This indeterminate limit cannot be simplified with algebra. So what can we
0
do?
Let’s approach the problem graphically. The graphs of f (x) = x 2 and g(x) = 2 x − 1 are
shown next at the left. Since we are interested in a limit as x → 0, we should focus on what
happens as we look at smaller and smaller graphing windows around x = 0, as shown in
the second and third graphs.
f (x) = x 2 and g(x) = 2 x − 1
Same graph but in smaller window
Graphs are almost linear here
y
y
y
2
1
⫺2
⫺1
1
2
x
0.4
0.08
0.2
0.04
⫺0.50 ⫺0.25
⫺1
0.25
0.50
x
⫺0.10 ⫺0.05
⫺0.2
⫺0.04
⫺0.4
⫺0.08
0.05
0.10
x
Near x = 0, the graph of f (x) = x 2 looks a lot like its horizontal tangent line y = 0, and
the graph of g(x) = 2 x − 1 looks a lot like its tangent line y = x. Thus we would anticipate
the behavior of the quotient
f (x)
g(x)
=
x2
−1
2x
as x → 0 to be similar to that of the quotient
0
x
of the corresponding tangent lines at x = 0. From this information it would be reasonable
x2
x −1
2
x→0
to guess that lim
= 0. As we are about to see, that is in fact the case. Indeed, we will
0
soon see that, in general, limits of the indeterminate form are related to the limit of the
0
quotient of the slopes, or derivatives, of those numerator and denominator functions.
CONFIRMING PAGES
0:9
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.6
We can also use tangent lines to examine limits of the indeterminate form
ample, consider the limit
x2
lim
.
x→∞ 2 x − 1
This is the same quotient
x2
2x − 1
303
L’Hôpital’s Rule
∞
.
∞
For ex-
of functions as before, but with a limit as x → ∞ instead
of as x → 0. As x → ∞ we have x 2 → ∞ and 2 x − 1 → ∞, and therefore this limit is of
∞
the indeterminate form .
∞
Again, we cannot simplify this expression with algebra. Let’s examine what happens as
we look at larger graphing windows to see the behavior of f (x) = x 2 and g(x) = 2 x − 1 as
x → ∞:
f (x) = x 2 and g(x) = 2 x − 1
Same graph but in larger window
y
y
3000
40
30
1
1
4000
50
1
1
y
60
2
2
Graphs are much different out here
2
2000
20
x
1000
10
2
2
4
6
x
2
2
4
6
8
10 12
x
From the rightmost graph we can see that the heights on the graph of g(x) = 2 x − 1 are in
some sense approaching ∞ in a fundamentally faster way than the heights on the graph of
f (x) = x 2 . Specifically, the slopes of g(x) = 2 x −1 are significantly steeper than the slopes of
f (x) = x 2 for large values of x. Accordingly, we would anticipate that as x → ∞ the values
x2
x −1
2
x→∞
of 2 x − 1 would win the race to ∞ and that the limit lim
would therefore be equal
to zero. Once again a ratio of slopes has helped us to guess the value of the limit of a
quotient. This geometric intuition is the basis for the powerful limit technique known as
L’Hôpital’s rule.
L’Hôpital’s Rule for the Indeterminate Forms
0
0
and
∞
∞
The ideas we have developed regarding limits of quotients suggest the theorem that fol0
lows. This key theorem will allow us to solve some limits of the indeterminate form or
∞
∞
THEOREM 3.14
0
by relating them to the quotients of the corresponding derivatives:
L’Hôpital’s Rule
Suppose f and g are differentiable functions on some punctured interval around x = c
f (x)
0
∞
on which g(x) is nonzero. If lim
is of the indeterminate form or , then
x→c g(x)
0
∞
f (x)
f (x)
lim
= lim ,
x→c g(x)
x→c g (x)
as long as the second limit exists or is infinite.
The conclusion holds also if x → ∞ (or x → −∞), as long as f and g are differentiable
on some interval (N, ∞) (or (−∞, N )) on which g(x) is nonzero.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
304
Chapter 3
December 11, 2012
Applications of the Derivative
There are a few technical points to notice about the hypotheses of this theorem. First, notice
that the functions f and g must be differentiable with g(x) nonzero for x-values near, but
not necessarily at, x = c. Second, notice that the conclusion does not necessarily hold if
the limit on the right does not exist for some reason other than being infinite. In practice,
most of the functions we look at will satisfy the two conditions we just mentioned. The
vital thing to check when applying L’Hôpital’s rule is that the limit is of the indeterminate
0
∞
form or . If the limit is not in one of those two indeterminate forms, then L’Hôpital’s
0
∞
rule cannot be applied.
The proof of L’Hôpital’s rule requires a more general version of the Mean Value
Theorem called the Cauchy Mean Value Theorem. This theorem—and the resulting proof
of L’Hôpital’s rule—are not that difficult, but proper proofs would take us too far afield for
our purposes, so we leave that for future mathematics courses. To simplify matters we will
prove L’Hôpital’s rule only in the special case where f and g are very well behaved and the
0
0
limit is of the indeterminate form :
Proof. We prove the theorem in the special case where f and g are continuous and differentiable
on an interval around and including x = c, and f (c) = g(c) = 0 but g (c) = 0. We wish to show that
lim
x→c
f (x)
f (c)
= .
g(x)
g (c)
We will work backwards from the right-hand side of the equality to the left-hand side. Applying
the definition of the derivative to the numerator and the denominator, we have
f (x) − f (c)
lim
f (c)
f (x) − f (c)
f (x) − 0
f (x)
x→c
x−c
=
= lim
= lim
.
= lim
g(x) − g(c)
x→c g(x) − g(c)
x→c g(x) − 0
x→c g(x)
g (c)
lim
x→c
x−c
L’Hôpital’s rule can be an extremely powerful tool for resolving indeterminate limits of
0
∞
the indeterminate form or . In Example 2 we will use the rule to find the limits that
0
∞
we examined graphically at the start of this section. L’Hôpital’s rule can also be useful for
resolving other indeterminate forms, provided that we can rewrite them so that they are of
0
∞
the form or ; see Example 3.
0
∞
Using Logarithms for the Indeterminate Forms 00 , 1∞ , and ∞0
Recall from Section 1.6 that limits of the form 00 , 1∞ , and ∞0 are indeterminate. For
example, all three of the following limits are of the indeterminate form 1∞ , but each one
of them approaches something different.
x2
x
= ∞;
x→∞ x − 1
lim
lim+ x1/(x
x→1
2
−1)
=
√
e;
1 x
lim 1 +
= e.
x→∞
x
In each of these limits there is a race between how fast the base approaches 1 and how
fast the exponent approaches ∞, and in some sense the winner of that race determines the
limit. But how can we determine who wins this race?
One difficulty with limits of the indeterminate forms 00 , 1∞ , and ∞0 is that such limits
involve a variable in both the base and the exponent. Fortunately, logarithms have the power
to change exponentiation into multiplication, in the sense that ln(ab ) = b ln a. The key to
using logarithms to calculate limits is the following theorem:
CONFIRMING PAGES
0:11
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.6
THEOREM 3.15
L’Hôpital’s Rule
305
Relating the Limit of a Function to the Limit of Its Logarithm
(a) If lim ln( f (x)) = L , then lim f (x) = e L .
x→c
x→c
(b) If lim ln( f (x)) = ∞ , then lim f (x) = ∞.
x→c
x→c
(c) If lim ln( f (x)) = −∞ , then lim f (x) = 0.
x→c
x→c
To use this theorem to calculate a limit of the form lim u(x)v(x) , we consider instead the limit
x→c
lim ln(u(x)v(x) ) = lim v(x) ln(u(x)).
x→c
x→c
Notice that the logarithm allows us to consider the limit of a product rather than a limit
involving an exponent. Once we find this limit L, the theorem tells us that the answer to
our original limit lim u(x)v(x) must be e L . If instead of L we get ±∞, our original limit must
x→c
be equal to e ±∞ .
Proof. We will prove only the first part of the theorem. The proof will follow directly from the
composition rule for limits of continuous functions from Section 1.5. Suppose f is a function that is
positive as x approaches c, so that ln( f (x)) is defined near x = c. For functions f with an exponent
and base both involving the variable x, domain restrictions will ensure that this will always be the
case. Since the function f (x) = ln x is continuous on (0, ∞), by the rule for limits of compositions
of continuous functions, we have
L = lim ln( f (x)) = ln( lim f (x)).
x→c
x→c
Since L = ln(A) if and only if A = e L , this equation implies that, as desired,
lim f (x) = e L .
x→c
Examples and Explorations
EXAMPLE 1
Checking to see if L’Hôpital’s rule applies
Determine whether or not L’Hôpital’s rule applies to each of the following limits as they
are written here (without any preliminary algebra or simplification):
(x − 1)2
x→1 x − 1
(a) lim
x2
x→0 1 + 2 x
(b) lim
2x
x→∞ 1 − 3 x
(c) lim (2 x − x 3 )
(d) lim
x→∞
SOLUTION
(a) L’Hôpital’s rule does apply, since as x → 1, both the numerator (x−1)2 and the denom(x−1)2
x→1 x−1
inator x − 1 approach zero, and therefore the limit lim
0
.
0
is of the indeterminate
form The more technical hypotheses of L’Hôpital’s rule are also satisfied by the numerator and denominator, as they are in most common examples. It is also possible to
use algebra to solve this limit.
(b) L’Hôpital’s rule does not apply, since
(c) L’Hôpital’s rule does not apply, since
x2
02
0
approaches
=
1 + 2x
1 + 20
2
x
3
2 − x is not a quotient.
= 0 as x → 0.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
306
Chapter 3
December 11, 2012
Applications of the Derivative
(d) L’Hôpital’s rule does apply, since as x → ∞ we have 2 x → ∞ and 1 − 3 x → −∞ and
∞
therefore this limit is of the indeterminate form
. The negative sign in the denomi−∞
nator could be factored out, but L’Hôpital’s rule will work even if we do not extract the
negative sign.
EXAMPLE 2
Applying L’Hôpital’s rule
Use L’Hôpital’s rule to calculate
SOLUTION
(a) Since the limit lim
x2
x −1
2
x→0
x2
x→0 2 x − 1
(a) lim
x2
.
x→∞ 2 x − 1
and (b) lim
0
0
is of the indeterminate form , L’Hôpital’s rule applies and
says that we can calculate it by considering instead the limit of the quotient of the
derivatives of the numerator and denominator:
d
(x 2 )
2(0)
0
x2
2x
L H
dx
= lim
=
lim
=
=
= 0.
0
x→0 2 x − 1
x→0 d
x→0 ( ln 2)2 x
(
ln
2)2
ln
2
x
(2 − 1)
lim
dx
Notice that we wrote “L’H” above the equals sign where we applied L’Hôpital’s rule, to
indicate our reasoning in that step. Notice also that we did not apply the quotient rule
f (x)
to differentiate the quotient
, because that is not the way that L’Hopital’s rule works.
g(x)
Instead, following L’Hôpital’s rule, we differentiated the numerator and denominator
individually.
x2
x −1
2
x→∞
(b) The limit lim
is of the indeterminate form
∞
,
∞
so L’Hôpital’s rule applies. Again
we replace the numerator and denominator of our quotient with their derivatives:
d
(x 2 )
x2
2x
L H
dx
= lim
= lim
.
lim x
x→∞ 2 − 1
x→∞ d
x→∞ ( ln 2)2 x
x
(2 − 1)
dx
Unfortunately, our application of L’Hôpital’s rule was not sufficient to resolve the limit
here, because if we let x → ∞ we have 2x → ∞ and ( ln 2)2 x → ∞, so our limit is still
∞
in the indeterminate form . However, this means that we can apply L’Hôpital’s rule
∞
again!
x2
2x
2
L H
L H
= lim
= lim
x→∞ 2 x − 1
x→∞ ( ln 2)2 x
x→∞ ( ln 2)( ln 2)2 x
lim
As x → ∞ the denominator of this limit approaches ∞, while the numerator is equal
to 2. Therefore we have
2
lim
= 0.
x→∞ ( ln 2)( ln 2)2 x
Notice that both of the preceding answers agree with what we guessed from a graphical analysis at the start of this section.
EXAMPLE 3
Rewriting limits in the form
0
0
or
∞
∞
so that L’Hôpital’s rule applies
Use L’Hôpital’s rule to calculate each of the following limits:
(a) lim x 2 e −3x
x→∞
(b) lim+ x ln x
x→0
1
1
−
sin x
x→0 x
(c) lim
CONFIRMING PAGES
0:13
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.6
307
L’Hôpital’s Rule
SOLUTION
(a) As x → ∞ we have x 2 → ∞ and e−3x → 0, so the limit lim x 2 e−3x is in the indeterx→∞
minate form ∞ · 0. Therefore, this limit is not yet in a form to which we can apply
L’Hôpital’s rule. Luckily, limits of the indeterminate form ∞ · 0 can always be rewrit∞
0
ten as a quotient of the form
or as a quotient of the form , simply by inverting
∞
0
one of the factors and placing it in the denominator. We can then choose whichever
of these two indeterminate forms we prefer and apply L’Hôpital’s rule. One way we
could rewrite the limit is
x2
x2
= lim 3x ,
−3x
x→∞ 1/e
x→∞ e
lim x 2 e −3x = lim
x→∞
which is of the indeterminate form
∞
.
∞
Another way we could write the limit is
e −3x
,
x→∞ 1/x 2
lim x 2 e −3x = lim
x→∞
0
0
which is of the indeterminate form .
The first way of rewriting seems like it would be easier to deal with, so we apply
L’Hôpital’s rule to that version:
x2
x→∞ e 3x
2x
L H
= lim
x→∞ 3e 3x
2
L H
= lim
x→∞ 9e 3x
lim x 2 e −3x = lim
x→∞
← rewrite to form
∞
∞
← apply L’Hôpital’s rule; still of form
∞
∞
← apply L’Hôpital’s rule
← since 9e 3x → ∞ as x → ∞
= 0.
(b) Note that we consider only the limit from the right, since ln x is not defined for negative
numbers. As x → 0+ we have x → 0 and ln x → −∞, and therefore the limit lim+ x ln x
x→0
is in the indeterminate form 0(−∞). In order to apply L’Hôpital’s rule we must rewrite
0
∞
the limit in the form or . In this case it is easier to leave ln x in the numerator; as
0
∞
+
x → 0 we have
ln x
x→0 1/x
lim+ x ln x = lim+
x→0
L H
← limit is now in the form
= lim+
1/x
−1/x 2
← apply L’Hôpital’s rule
= lim+
−x 2
x
← use algebra to simplify
x→0
x→0
= lim+ (−x) = 0.
x→0
−∞
∞
← simplify more and evaluate limit
Note that we could have applied L’Hôpital’s rule a second time in this problem,
since immediately after the first application of the rule the limit was again in the in∞
determinate form . However, simplifying instead resulted in a very simple limit that
∞
we could easily evaluate.
(c) As x → 0, both
1
x
and
1
sinx
become infinite, so the limit lim
of the indeterminate form ∞ − ∞.
x→0
1
x
−
CONFIRMING PAGES
1
sinx
is potentially
20:18
TKmaster2010
WHF00153/FREE087-Taalman
308
Chapter 3
December 11, 2012
Applications of the Derivative
To be honest, we are playing pretty fast and loose here: Specifically, we are not both1
1
ering to examine whether the terms and
approach ∞ or −∞. Without knowing
x
sinx
that, we don’t know for sure whether this is a limit of the indeterminate form ∞ − ∞
or the non-indeterminate form ∞ + ∞. We could look from the left and right and do
this more precisely if we cared to, but it is easier to instead do some algebra so that we
can apply L’Hôpital’s rule:
1
1
−
sin x
x→0 x
lim
= lim
x→0
sin x − x
x sin x
cos x − 1
L H
= lim
x→0 sin x + x cos x
− sin x
L H
= lim
x→0 cos x + cos x − x sin x
CHECKING
THE ANSWER
← combine fractions; form is now
0
0
← apply L’Hôpital’s rule; still form
0
0
← apply L’Hôpital’s rule again
=
− sin 0
cos 0 + cos 0 − 0 sin 0
← evaluate limit
=
0
= 0.
2
← use trigonometric values
All three limits we just calculated happened to be equal to zero, the first as x → ∞ and the
last two as x → 0. We can check these limits with calculator graphs:
lim x 2 e −3x = 0
lim+ x ln x = 0
lim
x→0
x→∞
x→0
1
1
−
x
sin x
1.5
0.08
=0
4
⫺3
⫺1
4
⫺1.5
0
EXAMPLE 4
3
2
0
⫺4
Using logarithms to calculate a limit
Use logarithms and L’Hôpital’s rule to calculate each of the following limits:
(a)
lim x 1/x
(b)
x→∞
SOLUTION
(a) Since x → ∞ and
∞0 .
1
x
lim (sin x) x
x→0+
→ 0 as x approaches ∞, this limit is of the indeterminate form
Let’s calculate the related, but different, limit lim ln(x1/x ) and see what we get:
x→∞
lim ln(x1/x ) = lim
1
x→∞ x
x→∞
= lim
ln x
ln x
x→∞ x
L H
= lim
x→∞
= lim
1/x
1
1
x→∞ x
= 0.
← algebra; now the limit is of the form 0 · ∞
← algebra; now the limit is of the form
← apply L’Hôpital’s rule
← simplify and evaluate the limit
CONFIRMING PAGES
∞
∞
0:15
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.6
309
L’Hôpital’s Rule
But wait! This is not the answer to our original limit. Since this limit is equal to 0, we
now know by Theorem 3.15 that our original limit must be equal to e 0 = 1.
(b) Once again we have a limit that involves a variable in both the base and the exponent, and we will need to use logarithms to resolve this problem. As x → 0+ we have
sin x → 0 and x → 0, so lim+ (sin x) x is in the indeterminate form 00 . Let’s look at the
x→0
related limit obtained by first taking the logarithm:
lim ln((sin x) x ) = lim+ x ln(sin x)
x→0+
x→0
← algebra; limit is now of form 0(−∞)
ln(sin x)
1/x
← algebra; limit is now of form
= lim+
(cos x)/(sin x)
−1/x 2
← apply L’Hôpital’s rule
= lim+
−x 2 cos x
sin x
← algebra; limit is now of form
−2x cos x + x 2 sin x
cos x
← apply L’Hôpital’s rule again
= lim+
x→0
L H
x→0
x→0
L H
= lim+
x→0
=
0+0
= 0.
1
−∞
∞
0
0
← evaluate the limit
Since this limit is equal to 0, our original limit lim+ (sin x) x must be equal to e 0 = 1.
x→0
What sorts of limits does L’Hôpital’s rule help us calculate?
TEST YOUR
? UNDERSTANDING
How do you explain in words what L’Hôpital’s rule says we can do to solve limits?
When calculating a limit by using L’Hôpital’s rule multiple times, how do you know
when to stop applying L’Hôpital’s rule and evaluate the limit?
How can L’Hôpital’s rule sometimes be used to solve limits of the indeterminate form
0 · ∞?
How can we use logarithms to solve limits of the indeterminate forms 00 , 1∞ , and ∞0 ?
EXERCISES 3.6
Thinking Back
Indeterminate forms: Determine which of the given forms
are indeterminate. For each form that is not indeterminate,
describe the behavior of a limit of that form.
∞·∞
0
0
0∞
∞
∞
00
∞∞
1∞
∞−∞
∞·0
∞
0
0
∞
∞0
Simple limit calculations: Determine each of the limits that
follow. You should be able to solve all of these very quickly
by thinking about the graphs of the functions.
lim 2 x
x→∞
1
x2
lim tan x
lim
x→0
x→π/2
lim x −5
lim ln x
x→∞
x→0+
lim sin x
x→∞
CONFIRMING PAGES
lim e 3x
x→−∞
x
lim
x→∞
1
2
lim log1/2 x
x→∞
20:18
TKmaster2010
WHF00153/FREE087-Taalman
310
Chapter 3
November 24, 2012
Applications of the Derivative
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If a limit has an indeterminate form, then
that limit does not have a real number as its solution.
(b) True or False: L’Hôpital’s rule can be used to find the
limit of any quotient
f (x)
as x → c.
g(x)
(c) True or False: When using L’Hôpital’s rule, you need
to apply the quotient rule in the differentiation step.
(d) True or False: L’Hôpital’s rule applies only to limits as
x → 0 or as x → ∞.
(e) True or False: L’Hôpital’s rule applies only to limits of
the indeterminate form
4. Explain graphically why it might make sense that a limit
f (x)
∞
of the indeterminate form
would be related
g(x)
∞
f (x)
to lim .
x→∞ g (x)
lim
x→∞
5. Suppose f (x) = x 2 − 1 and g(x) = ln x. Find the equations of the tangent lines to these functions at x = 1.
Then argue graphically that it would be reasonable to
think that the limit of the quotient
equal to the limit of the quotient of these tangent lines as
x → 1.
6. Suppose f (x) = 2 x − 4 and g(x) = x − 2. Find the equations of the tangent lines to these functions at x = 2. Then
argue graphically that it would be reasonable to think
0
∞
or .
0
∞
that the limit of the quotient
x→2
ln 4.
(g) True or False: If lim ln( f (x)) = ∞, then lim f (x) =
x→2
x→2
∞.
(h) True or False: If lim ln( f (x)) = −∞, then lim f (x) =
x→2
x→2
−∞.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) Three limits of the indeterminate form
f (x)
as x → 2 might be
g(x)
equal to the limit of the quotient of these tangent lines as
x → 2.
(f) True or False: If lim ln( f (x)) = 4, then lim f (x) =
x→2
f (x)
as x → 1 might be
g(x)
Each of the limits in Exercises 7–12 is of the indeterminate
form 0 · ∞ or ∞ · 0. Rewrite each limit so that it is (a) in
0
∞
and then (b) in the form . Then (c) determine
0
∞
the form
which of these indeterminate forms would be easier to work
with when applying L’Hôpital’s rule.
7.
lim 2−x x
8. lim(2 x − 1)x −2
x→∞
lim x
−2
x→0
ln x
10. lim x 3 ln x
0
, one that
0
9.
that
11.
that
that
13. Find the error in the following incorrect calculation, and
then calculate the limit correctly:
approaches ∞, one that equals 0, and one
equals 3.
∞
(b) Three limits of the indeterminate form , one
∞
approaches ∞, one that equals 0, and one
equals 3.
(c) Three limits of the indeterminate form 0 · ∞,
that approaches ∞, one that equals 0, and one
equals 3.
x→∞
lim x csc x
x→0
lim
one
that
3. Explain graphically why it makes sense that a limit
f (x)
0
of the indeterminate form would be related to
x→c g(x)
0
f (x)
lim .
x→c g (x)
lim
x→0
x→0
12. lim
x→1
√
x − 1 ln(x − 1)
x 2 − x L H
2x − 1 L H
2
= lim
= lim
x→0 ( ln 2)2 x
x→0 ( ln 2)2 2 x
2x − 1
2
2
=
=
.
( ln 2)2 20
( ln 2)2
14. Find the error in the following incorrect calculation. Then
calculate the limit correctly.
lim
x→∞
e −x L H
−e −x L H
e −x
0
= = 0.
= lim
= lim
2
x→∞
x→∞
x
2x
2
2
Skills
Calculate each of the limits in Exercises 15–20 (a) using
L’Hôpital’s rule and (b) without using L’Hôpital’s rule.
x2 + x − 2
x→1
x−1
x−1
17. lim
x→∞ 2 − 3x 2
15. lim
19.
lim
x→∞
e 3x
1 − e 2x
x 2 − 4x + 4
x→2
x−2
3x
18. lim
x→∞ 1 − 4 x
2x − 1
20. lim x
x→0 4 − 1
16. lim
Calculate each of the limits in Exercises 21–48. Some of these
limits are made easier by L’Hôpital’s rule, and some are not.
21. lim
x→3
2x − 8
3−x
e 2x−4 − 1
x2 − 4
2−x
25. lim 2
x→∞ x + 1
23. lim
x→2
x−1
e x−1 − 1
ex − 7
24. lim
x→∞ 8x 2 + 12x + 5
22. lim
x→1
26.
lim
x→∞
CONFIRMING PAGES
x + 3x
3 − 2x
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
3.6
x −2
+1
ex
lim e x −
27. lim
28.
29.
30.
x→0 x −3
31.
33.
35.
37.
39.
41.
43.
45.
x+1
x→∞
x
1
2
32.
xe x
e 2x + 1
34.
lim x
x→∞
lim
x→∞
lim x 2 e −x
lim+ x −1 −
x→∞
x→0
x→∞ x 2
e −3x
x→∞ x 2 + 3x + 1
log2 x
lim
x→0+ log2 3x
ln(x − 2)
lim
x→2+ ln(x 2 − 4)
cos x − 1
lim
x→0
sin x
1 − cos x
lim
x→0
tan x
x cos x
lim
x→0 1 − e x
lim+
x→0
Intuitively, f dominates g as x → ∞ if f (x) is very much larger
than g(x) for very large values of x. In Exercises 65–74, use
limits to determine whether u(x) dominates v(x), or v(x)
dominates u(x), or neither.
1
2 −1
x
xe −3x
+ 3x + 1
lim
2x − 1
1
−
x2
1 − ex
65. u(x) = x + 100, v(x) = x
ln x
x→∞ ln(2x + 1)
x
2 −4
38. lim+ ln
lim
36.
lim
x−2
x→2
ln x
x→0 ln x − x + 1
sin x
42. lim
x→0 x + sin x
sin(cos x)
44. lim
x→π/2
cos x
sin( ln x)
46. lim
x→1 x − 1
lim+
40.
−1
−1
tan x
tan x
48. lim
x→0 sin−1 x
sin x
Calculate each of the limits in Exercises 49–64. Some of these
limits are made easier by considering the logarithm of the
limit first, and some are not.
47. lim
x→0
49.
51.
53.
55.
57.
59.
61.
63.
lim x ln x
50.
x→0+
lim+ (x − 2) x
2
−4
52.
x→2
lim+ x 1/(x−1)
54.
x→1
lim x 1/x
x→∞
lim
x→∞
1
x+1
56.
x
58.
lim+ (x 2 + 1) x
x→0
lim x2x
x
1
lim
x→0+
x→∞
lim
x→∞
x
x−1
x
lim+ x sin x
62.
lim+ (cos x)1/x
64. lim(1 − cos x) x
lim+ ( ln x) x−1
x→1
lim+ (sin 3x)2x
x→0
x→0
x→0
A function f dominates another function g as x → ∞ if f (x)
and g(x) both grow without bound as x → ∞ and if
lim
x→∞
67. u(x) = 100x 2 , v(x) = 2x100
68. u(x) = x 2 , v(x) = 2 x
69. u(x) = 2 x , v(x) = e x
70. u(x) = x10 + 3, v(x) = 10 x + 3
71. u(x) = log2 x, v(x) = log30 x
72. u(x) = ln(x 2 + 1), v(x) = x 2 + 1
73. u(x) = 0.001e 0.001x , v(x) = 100x100
74. u(x) = 0.001x 2 − 100x, v(x) = 100 log3 x
As you will prove in Exercises 93 and 94, exponential growth
functions e kx always dominate power functions x r , and power
functions x r with positive powers always dominate logarithmic functions logb x. Use these facts to quickly determine each
of the limits in Exercises 75-80.
x101 + 500
x→∞
ex
√
x
77. lim
x→∞ 300 ln x
75.
79.
lim
lim 2 x x −100
76.
78.
80.
x→∞
lim
x→∞ x101
lim
x→∞
ex
+ 500
ln x100
x8
lim e −x ln x
x→∞
Now that we know L’Hôpital’s rule, we can apply it to solve
more sophisticated global optimization problems. Consider
domains, limits, derivatives, and values to determine the
global extrema of each function f in Exercises 81–86 on the
given intervals I and J.
81. f (x) = x ln x, I = (0, 1], J = (0, ∞)
x
60.
x→0
66. u(x) = 5x 2 + 1, v(x) = x 3
lim x ln x
x→∞
lim+ (x − 1)ln x
x→1
311
L’Hôpital’s Rule
82. f (x) = x 2 ln(0.2x), I = (0, 4], J = (0, ∞)
83. f (x) = x 3 e −x , I = [0, ∞), J = (−∞, ∞)
ln x
, I = [0, 1], J = [1, ∞)
ln 2x
sin x
, I = (0, π ], J = (0, 2π )
85. f (x) =
1 − cos x
ex
86. f (x) =
, I = [0, ∞), J = (−∞, ∞)
1 + x2
84. f (x) =
f (x)
= ∞.
g(x)
Applications
In Exercises 87–89, suppose that Leila is a population
biologist with the Idaho Fish and Game Service. Wolves were
introduced formally into Idaho in 1994, but there were some
wolves in the state before that. Leila has been assigned the
task of estimating the rate at which the number of wolves
in Idaho increased naturally, before the animals were introduced. The only information she has is a population
model, which indicates that the wolf population currently
satisfies the formula
w(t) = 835(1 − e −.006t ),
where t is the number of years since 1994.
87. Leila reasons that the average rate of change of wolves
w(t)
. Why does her
from 1994 to a given time t is given by
t
reasoning make sense?
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
312
Chapter 3
November 24, 2012
Applications of the Derivative
models approximate the number of elk over time as
88. To approximate the rate of increase of wolves per year at
the beginning of 1994, Leila decides to take the limit of
w(t)
as t → 0. Why does this approach make sense, and
t
E(t) =
what is the value of that limit? Is there another way she
could find the same number?
E0 + 72e −0.006t sin(πt/4) + 8te −0.006t sin(πt/4)
1 + 0.2W0
(a) Use the Squeeze Theorem for Limits to show that the
89. Leila must also determine hunting policies to sustain a
population W0 of wolves that satisfy federal guidelines,
while maximizing the sustained elk population E0 for
which the state can sell hunting tags. Her predator-prey
population goes toward
E0
as t → ∞.
1+0.2W0
(b) Explain why L’Hôpital’s Rule is not a good method for
calculating this limit.
Proofs
90. In your own words, prove the special case of L’Hôpital’s
rule that is proved in the reading. Explain each step in
detail.
Exercises 91–94 concern dominance of functions as defined
earlier in Exercises 65–74.
93. Use L’Hôpital’s rule to prove that exponential growth
functions always dominate power functions.
94. Use L’Hôpital’s rule to prove that power functions with
positive powers always dominate logarithmic functions.
95. Use logarithms to prove that for any real number r,
91. Use L’Hôpital’s rule to prove that every exponential growth
function dominates the power function g(x) = x 2 .
92. Use L’Hôpital’s rule to prove that every power function
with a positive power dominates the logarithmic function
g(x) = ln x.
lim 1 +
x→∞
r x
= e r.
x
Thinking Forward
Convergence and divergence of sequences: If a sequence
a1 , a2 , a3 , . . . , ak , . . . approaches a real-number limit as
k → ∞, then the sequence {ak } converges. If the terms of the
sequence do not get arbitrarily close to some real number,
then the sequence diverges. Determine the general form {ak }
for each of the following sequences, and then use L’Hôpital’s
rule to determine whether that sequence converges or
diverges.
1 4 9 16 25 36
, , ,
,
,
...
2 4 8 16 32 64
ln 1 ln 2 ln 3 ln 4 ln 5 ln 6
,
,
,
,
,
...
ln 2 ln 3 ln 4 ln 5 ln 6 ln 7
3
7
15
31
63
1
,
,
,
,
,
, ...
10 100 1000 10,000 100,000 1,000,000
8
27 64 125 216
1
,
,
,
,
,
, ...
301 304 309 316 325 336
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 24, 2012
Chapter Review, Self-Test, and Capstones
313
CHAPTER REVIEW, SELF-TEST, AND CAPSTONES
Before you progress to the next chapter, be sure you are familiar with the definitions, concepts, and basic skills outlined here.
The capstone exercises at the end bring together ideas from this chapter and look forward to future chapters.
Definitions
Give precise mathematical definitions or descriptions of each
of the concepts that follow. Then illustrate the definition with
a graph or algebraic example, if possible.
f has a local maximum or a local minimum at x = c
f has a critical point at x = c
f is increasing or decreasing on an interval I
f is concave up or concave down on on interval I
f has a global maximum or a global minimum at x = c
f does not have any global maximum or does not have
any global minimum
The first-derivative test: Suppose x = c is a
of a differ, then f has a local maximum
entiable function f . If
, then f has a local minimum at x = c.
at x = c. If
, then f has neither a local maximum nor a local
If
minimum at x = c.
Theorems
Fill in the blanks to complete each of the following theorem
statements:
If x = c is a local extremum of f , then f (c) is either
.
Rolle’s Theorem: If f is
on [a, b] and
on (a, b),
, then there exists at least one value c ∈ (a, b)
and if
.
for which f (c) =
Suppose f and f are differentiable on an interval I. If
is positive on I, then f is concave up on I. If
is
on I, then f is concave down on I.
The Mean Value Theorem: If f is
on [a, b] and
on
(a, b), then there exists at least one value c ∈ (a, b) for
.
which f (c) =
If f is differentiable on an interval I and f is positive in
on I.
the interior of I, then f is
The second-derivative test: Suppose x = c is a
of a
, then f has a local
twice-differentiable function f . If
, then f has a local minimum
maximum at x = c. If
, then this test is inconclusive.
at x = c. If
L’Hôpital’s Rule: If f and g are
or
If f is differentiable on an interval I and f is
interior of I, then f is decreasing on I.
If f is differentiable on an interval I and f is zero in the
on I.
interior of I, then f is
near x
in the
If f (x) = g (x) for all x ∈ [a, b], then for some constant C,
for all x ∈ [a, b].
f (x) =
form
and g(x) is nonzero
f (x)
= c, and if lim
is of the indeterminate
x→c g(x)
f (x)
or
, then lim
=
.
x→c g(x)
If lim ln( f (x)) = L, then lim f (x) =
If lim ln( f (x)) =
, then lim f (x) = ∞.
If lim ln( f (x)) =
, then lim f (x) = 0.
x→c
x→c
x→c
x→c
.
x→c
x→c
Geometric Formulas and Theorems
Volume and Surface Area Formulas: Write a formula for
(a) the volume and (b) the surface area of each solid described
below.
A rectangular box of width w, length l, and height h
A box of height h with a square base of area A
A sphere of radius r
A sphere of circumference C
A right circular cylinder of radius R and height y
A right circular cylinder whose radius r is half of its
height
A cone of height y and circular base of area A
Right-Triangle Theorems: Write an equation that describes the
relationships between the variables given for each theorem
that follows. Draw a picture to illustrate the theorem and the
roles of the variables in your equation.
The Pythagorean Theorem: If a right triangle has legs of
.
lengths x and y and a hypotenuse of length h, then
The Law of Similar Triangles: Suppose two right triangles
have the same angle measures as each other (i.e., they are
similar), where the first has legs of lengths x1 and y1 and a
hypotenuse of length h1 and the second has corresponding legs of lengths x2 and y2 and a hypotenuse of length
h2 . Then we have the following three equations involving
,
, and
.
ratios:
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
314
Chapter 3
November 24, 2012
Applications of the Derivative
Skill Certification: Curve Sketching and L’Hôpital’s Rule
Intervals of behavior: For each of the following functions f ,
determine the intervals on which f is positive, negative,
increasing, decreasing, concave up, and concave down.
1. f (x) = x 3 + 3x 2 − 9x − 27
2. f (x) = x4/3
x−1
3. f (x) =
x+3
1
4. f (x) =
(x − 1)(x + 2)
5. f (x) = e 1+x
2
7. f (x) = 2 x (2 x − 1)
6. f (x) = sin
11. f (x) =
x
x2 + 1
x
4
8. f (x) = sec2 x
13. f (x) = x ln x
15. f (x) =
ex
1 − ex
10. f (x) = x 3 − 15x 2 − 33x
12. f (x) =
√
19. f (x) = x x + 1
1
3x + 1
√
20. f (x) = x 2 + 2x + 10
21. f (x) = xe x
22. f (x) = ln(x 2 + 1)
23. f (x) = cos x
24. f (x) = tan−1 x
17. f (x) = x 3 − 2x 2 − 4x + 8
Important points: Find all roots, local maxima and minima,
and inflection points of each function f . In addition, determine whether any local extrema are also global extrema on the
domain of f .
9. f (x) = 3x 4 − 8x 3
Curve sketching: For each function f that follows, construct
sign charts for f , f , and f , if possible. Examine function
values or limits at any interesting values and at ±∞. Then
interpret this information to sketch a labeled graph of f .
1
√
1+ x
14. f (x) = x 4/3 − x1/3
16. f (x) = tan−1 x 2
18. f (x) =
L’Hôpital’s Rule limit calculations: Calculate each of the
limits that follow. Some of these limits are easier to calculate
by using L’Hôpital’s rule, and some are not.
x−2
+1
sin x
27. lim
x→0 cos x − 1
ln x
29. lim
x→∞ ln(x + 1)
25.
31.
lim
x→∞
26. lim
x−3
x→0
28.
3x − 1
2x − 1
lim x 3 e −x
x→∞
tan−1 x
1 − cos x
lim+ x1−cos x
30. lim
x→0
lim x1/(x−1)
32.
x→1+
x→0
Capstone Problems
A.
Critical points, extrema, and inflection points: Find examples
of differentiable functions which illustrate that not every
critical point is an extremum and that not every zero of
the second derivative is an inflection point. More specifically, find the following:
(a) A function f with f (2) = 0 and an extremum at
x = 2, and a function g with g (2) = 0 but no
extremum at x = 2.
(b) A function k with k (2) = 0 and an inflection point at
x = 2, and a function h with h (2) = 0 but no inflection point at x = 2.
B.
C.
Optimizing perimeter, given area: Suppose that you want
to cut a rectangular shape with a particular area A from
a sheet of material, and that you want the perimeter of the shape to be as small as possible. Use
techniques of optimization to argue that the smallest
possible perimeter will be achieved if the rectangular
shape that you cut out is a square.
The Mean Value Theorem: Recap the development of the
Mean Value Theorem as follows:
(a) Prove that if f is a differentiable function, then
every extremum x = c of f is also a critical point of f .
(Hint: Show that f (c) = 0 by proving that f + (c) ≤ 0
and f − (c) ≥ 0.)
(b) Use part (a) and the Extreme Value Theorem to prove
Rolle’s Theorem. (Hint: Consider the case where f has
an extremum on the interior of the interval first.)
(c) Explain how the Mean Value Theorem is essentially
a rotated version of Rolle’s Theorem and how the
proof in the reading makes use of that fact.
D. Area accumulation functions: Suppose f is the function pictured here, and A(x) is the associated function whose
value at any x ≥ 0 is equal to the area between the graph
of f and the x-axis from 0 to x. The quantity A(2.5) is
shaded in the figure. We will count area below the x-axis
negatively, so that in this example A(5) is less than A(4).
Area function A(x) is defined by
the shaded region as x varies
y
4
2
2
A(x)
1
2 x 3
4
5
x
4
6
(a) On what interval of x-values is A(x) an increasing
function? On what interval is A(x) decreasing?
(b) On what interval of x-values is the function f positive? On what interval is f negative?
(c) Here is a surprising fact: One of these functions is the
derivative of the other! Use your answers to parts (a)
and (b) to determine whether A = f or f = A.
CONFIRMING PAGES
20:18
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
C H A P T E R 4
Definite Integrals
4.1
Addition and Accumulation
Accumulation Functions
Sigma Notation
The Algebra of Sums in Sigma Notation
Formulas for Common Sums
Examples and Explorations
4.2
n
Riemann Sums
y
Subdivide, Approximate, and Add Up
Approximating Area with Rectangles
Riemann Sums
Signed Area
Types of Riemann Sums
Examples and Explorations
4.3
ak
k=1
x
Definite Integrals
Defining the Area Under a Curve
Properties of Definite Integrals
Formulas for Three Simple Definite Integrals
Examples and Explorations
4.4
Indefinite Integrals
Antiderivatives and Indefinite Integrals
Antidifferentiation Formulas
Antidifferentiating Combinations of Functions
Examples and Explorations
4.5
The Fundamental Theorem of Calculus
The Fundamental Theorem
Using the Fundamental Theorem of Calculus
The Net Change Theorem
The Proof of the Fundamental Theorem
Examples and Explorations
4.6
f (x) dx
b
f (x) dx = F(b) − F(a)
a
Areas and Average Values
The Absolute Area Between a Graph and the x-axis
Areas Between Curves
The Average Value of a Function on an Interval
The Mean Value Theorem for Integrals
Examples and Explorations
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
4.7
November 21, 2012
Functions Defined by Integrals
Area Accumulation Functions
The Second Fundamental Theorem of Calculus
Defining the Natural Logarithm Function with an Integral
The Proof of the Second Fundamental Theorem
Examples and Explorations
Chapter Review, Self-Test, and Capstones
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.1
4.1
317
Addition and Accumulation
ADDITION AND ACCUMULATION
An introduction to accumulation functions and the area under a curve
Sigma notation and properties of sums
A preview of limits of sums, the backbone of the definition of the definite integral
Accumulation Functions
With this chapter we begin the study of integrals. While derivatives describe the rates at
which functions change, integrals can describe how functions accumulate. As we will see
throughout this chapter, the concepts of accumulation, area, and differentiation are fundamentally intertwined.
For a simple example of how these concepts are intertwined, imagine that you are driving down a straight road that has stoplights. Suppose that you start from a full stop at one
stoplight, increase your velocity for 20 seconds to reach 60 miles per hour (88 feet per
second), then decrease your velocity for 20 seconds until you come to a full stop at the next
stoplight, as illustrated next at the left. Your speedometer works, so you can tell how fast
you are going at any time. However, your odometer is broken, so you have no idea how far
you have travelled. Using data from your speedometer, you can show that your velocity on
this trip is given by v(t) = −0.22t 2 + 8.8t feet per second, from t = 0 to t = 40 seconds, as
illustrated in the graph shown at the right.
v(t) = −0.22t 2 + 8.8t
0 mph
60 mph
0 mph
v
88
v 88 ft/s
t0
t 20
t 40
10
20
30
40
t
So, how far was it between the two stoplights? If you had travelled at a constant velocity,
this would be an easy application of the “distance equals average rate times time” formula.
For example, driving at exactly 30 miles per hour (44 feet per second) for 40 seconds would
accumulate a distance of d = (44)(40) = 1760 feet.
Unfortunately in our example, velocity varies. However, we can approximate the distance travelled by assuming a constant velocity over small chunks of time. For example,
we could use v(5), v(15), v(25), and v(35) as constant velocities over the intervals [0, 10],
[10, 20], [20, 30], and [30, 40], respectively. Using the d = rt formula over each of the four
time intervals would give us an approximate distance travelled of
d ≈ d 1 + d 2 + d 3 + d 4 = r 1 t 1 + r2 t 2 + r 3 t 3 + r 4 t 4
= v(5)(10) + v(15)(10) + v(25)(10) + v(35)(10)
= (38.5)(10) + (82.5)(10) + (82.5)(10) + (38.5)(10) = 2420 feet.
Notice that we have just estimated the distance between the two stoplights by means of
only the readings on your speedometer at t = 5, t = 15, t = 25, and t = 35. Despite
having used only this small amount of information, our estimate is fairly close to the actual
distance, which in this example happens to be just over 2346 feet.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
318
Chapter 4
November 21, 2012
Definite Integrals
We can think of the preceding four 10-second distance approximations as areas of four
rectangles, as shown next at the left. For example, over the first time interval we have a rectangle of width t 1 = 10, height r1 = v(5) = 38.5, and area d1 = (t 1 )(r1 ) = (38.5)(10) = 385.
Distance was approximated with
a sum of areas of rectangles
Thinner rectangles give a
better approximation
v
v
88.0
82.5
88.0
38.5
10
20
30
40
t
10
20
30
40
t
Notice that the sum of the areas of these rectangles is a pretty good approximation for the
distance between the stoplights and also a pretty good approximation for the area under
the velocity curve in the figure at the left. Using smaller time intervals would give us thinner
rectangles (as shown at the right), whose combined areas would be a better approximation for the distance travelled, as well as a better approximation for the area under v(t).
These figures suggest that the exact distance travelled might equal the exact area under the
velocity curve. Furthermore, since we already know that velocity is the derivative of
position, the figures also suggest connections between accumulation, area, and derivatives.
Over the rest of this chapter we will make these notions and connections more precise.
Sigma Notation
From the stoplight example we just discussed, it seems that we will have to do a lot of
adding in order to investigate area and accumulation functions. Our better approximation with smaller, but more, rectangles involved 16 rectangles with areas to calculate and
add up—and for an even better approximation we might consider 100 or more rectangles.
We now develop a compact notation to represent the sum of a sequence of numbers—in
particular, a sequence of numbers that has a recognizable pattern. This notation is called
sigma notation, since it uses the letter “sigma” (written ), which is the Greek counterpart
of the letter “S” (for “sum”).
DEFINITION 4.1
Sigma Notation
If a k is a function of k, and m and n are nonnegative integers with m ≤ n, then
n
a k = a m + a m+1 + a m+2 + · · · + a n−1 + a n .
k=m
As an initial example, consider our earlier sum of four distances. This sum can be written
in sigma notation as follows:
d≈
4
k=1
dk =
4
r k t k = 2420 feet.
k=1
The expression in Definition 4.1 is pronounced “the sum from k = m to n of a k .” The
“k = m” below the sigma shows where we should begin the sum, and it also specifies that
the index, or stepping variable, for the sum is k. With sigma notation, the step from one
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.1
Addition and Accumulation
319
value of k to the next is always equal to 1. If we start with k = 1, then the next value of k will
be k = 2, the next will be k = 3, and so on. We end at the value k = n, where n is the number appearing above the sigma in the notation. The terms a k represent the pattern of the
items to be summed. When we say that a k is a function of k, we mean that, for every integer
index k, the expression for a k determines a unique real number. Since the index k can have
only integer values, it is traditional to use a k rather than the usual function notation a(k).
When printed in a line of text, sigma notation is more compact and looks like this:
k=m a k . Sigma notation always follows the general pattern
n
ending value
(function of k).
k=starting value
To find the value of a sum in sigma notation nk=m a k you would evaluate the function a k
at k = m, k = m + 1, k = m + 2, and so on until k = n, and then add up all of these values.
To put a given sum into sigma notation you must identify a pattern for the function a k , as
well as starting and ending values for k. For example, we could represent the sum
1+
with the sigma notation
1
1
1
1
1
1
1
1
1
+ + + + + + + +
2
3
4
5
6
7
8
9
10
10
1
k=1 k ,
since each of the numbers in our sum is of the form
some integer k, with k = 1 for the first term and k = 10 for the last term.
1
k
for
The Algebra of Sums in Sigma Notation
The next theorem expresses two common properties of sums in sigma notation. For the
moment we will restrict our attention to sums whose index k starts at k = 1. However,
Theorem 4.2 is true for sums that begin at any integer k = m.
THEOREM 4.2
Constant-Multiple and Sum Rules for Sums
If a k and b k are functions defined for nonnegative integers k, and c is any real number,
then
n
n
n
n
n
ca k = c
a k.
(b)
(a k + b k ) =
ak +
b k.
(a)
k=1
k=1
k=1
k=1
k=1
Part (a) of this theorem is a general version of the distributive property c(x + y) = cx + cy.
Part (b) is a general version of the associative and commutative properties of addition (i.e.,
the fact that we can reorder and regroup the numbers in a sum). A simple example is the
fact that (a1 + b1 ) + (a 2 + b 2 ) = (a1 + a 2 ) + (b1 + b 2 ). The proof of Theorem 4.2 consists
mostly of translating the sigma notation.
Proof. To prove part (a), we translate the sigma notation to its expanded form and then apply the
distributive property:
n
ca k = ca1 + ca 2 + ca 3 + · · · + ca n−1 + ca n
← write out the sum
k=1
= c(a1 + a 2 + a 3 + · · · + a n−1 + a n )
=c
n
a k.
← factor out c
← write in sigma notation
k=1
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
320
Chapter 4
November 21, 2012
Definite Integrals
The proof of part (b) is similar; we will write the sigma notation in expanded form, reorder and
regroup the terms in the sum, and then write the reordered sum in sigma notation:
n
(a k + b k ) = (a1 + b1 ) + (a 2 + b 2 ) + (a 3 + b 3 ) + · · · (a n−1 + bn−1 ) + (a n + b n )
k=1
= (a1 + a 2 + a 3 + · · · + a n−1 + a n ) + (b1 + b 2 + b 3 + · · · + bn−1 + b n )
=
n
ak +
k=1
n
b k.
k=1
We can strip off terms from the beginning or end of a sum, or split a sum into two
pieces, by applying the following theorem:
THEOREM 4.3
Splitting a Sum
Given any function a k defined for nonnegative integers k, and given any integers m, n,
and p such that 0 ≤ m < p < n,
n
k=m
ak =
p−1
ak +
k=m
n
a k.
k=p
Proof. The proof of this theorem is fairly simple; we need only write out the terms of the sum. If
m < p < n, we have
n
a k = a m + a m+1 + · · · + ap−1 + ap + ap+1 + · · · + a n−1 + a n
← expand
k=m
= (a m + a m+1 + · · · + ap−1 ) + (ap + ap+1 + · · · + a n−1 + a n )
p−1
=
k=m
ak +
n
a k.
← regroup
← sigma notation
k=p
Formulas for Common Sums
2
It can be tedious to calculate a long sum by hand; for example, consider the sum 1000
k=1 k .
To calculate this sum directly, we would have to write out all 1000 of its terms and then add
them together. Luckily, the next theorem provides formulas for the values of a few simple
sums. Using these formulas, we can quickly calculate any sum whose general term is a
constant, linear, quadratic, or cubic polynomial in k.
THEOREM 4.4
Sum Formulas
If n is a positive integer, then
n
1=n
(a)
(b)
k=1
(c)
n
k=1
n
k=1
k2 =
n(n + 1)(2n + 1)
6
(d)
n
k=1
k=
n(n + 1)
2
k3 =
n2 (n + 1)2
4
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.1
Addition and Accumulation
321
Proof. To prove the first formula we only have to write out the sum
n
k=1
1 = 1 + 1 + 1 + · · · + 1 = n.
n times
You will prove the second formula by using two cases in Exercises 61 and 62. The third and
fourth formulas are simple to prove by a method called induction, but that method of proof is best
left for a future mathematics course, so we do not include proofs of the third and fourth formulas
at this time.
Examples and Explorations
EXAMPLE 1
Converting from sigma notation to an expanded sum
Write the sum represented by the sigma notation 8k=2 2k in expanded form, and find the
value of the sum.
SOLUTION
The given sigma notation is a compact way of writing “the sum of all numbers of the form
2k , where k is an integer greater than or equal to 2 and less than or equal to 8.” This means
we must find the values of 2k for k = 2, k = 3, and so on until k = 8, and then add all of
these values:
8
2k = 22 + 23 + 24 + 25 + 26 + 27 + 28 = 508.
k=2
EXAMPLE 2
Converting from an expanded sum to sigma notation
Write the following sum in sigma notation:
3
4
5
6
7
8
9
+ + + + + + .
4
5
6
7
8
9
10
SOLUTION
First, we look for a pattern in the numbers of the sum, so we can determine the function
k
a k . One possible pattern is that each number in the sum is of the form
for some nonnegative integer k. Therefore, we can set a k =
the value k = 3. The last number,
k
k+1
=
9
10
k
.
k+1
The first
k+1
3
number, , corresponds
4
9
, corresponds to the value k
10
= 9 when k = 9, we have
. The sum can be written in sigma notation as
9
k
3
4
5
6
7
8
9
+ + + + + +
=
.
4
5
6
7
8
9
10
k+1
k=3
EXAMPLE 3
to
Algebraically manipulating sums in sigma notation
Given that 4k=1 a k = 7 and 4k=1 b k = 10, find 4k=1 (a k + 3b k ).
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
322
Chapter 4
November 21, 2012
Definite Integrals
SOLUTION
Notice that we do not know what the functions a k and b k are in this problem, and thus we
cannot directly compute 4k=1 (a k + 3b k ). However, using Theorem 4.2, we have
4
4
4
(a k + 3b k ) =
ak +
3b k
k=1
k=1
=
4
ak + 3
4
k=1
bk
← Part (b) of Theorem 4.2
k=1
= 7 + 3(10) = 37.
EXAMPLE 4
← Part (a) of Theorem 4.2
k=1
← using the values of the sums given
Sums can be combined only if they start and end at the same value
√
19
1
Use the fact that 20
≈ 2.5977 and 19
k=2
k=0 k ≈ 57.1938 to estimate
k=2
k
2
k
+
√
k .
SOLUTION
We could of course just directly calculate the desired sum by adding up terms, but instead
we will combine the two given sums to obtain the information we need. The key is to
rewrite the two sums we were given so that they each begin and end and the same value;
we will rewrite so that they begin at k = 2 and end at k = 19.
1
1
1
1
1
We are given that 20
= + + · · · + + is approximately 2.5977. To get the
k=2
k
2
3
19
20
sum of the k = 2 through k = 19 terms, all we have to do is subtract the extra term
19
1
k=2
k
=
20
1
k=2
k
−
1
:
20
1
1
≈ 2.5977 −
= 2.5477.
20
20
√
Similarly, we can find the sum from k = 2 to k = 19 of k by subtracting the two extra
19 √
terms (at k = 0 and k = 1) at the beginning of the sum k=0 k :
19 √
19 √
√
√
k=
k − 0 − 1 ≈ 57.1938 − 0 − 1 = 56.1938.
k=2
k=0
We can now use Theorem 4.2 to combine these sums and find the value of the desired
sum:
19
2
k=2
k
+
√
k =
19
2
k=2
=2
k
+
← sum of sums
k=2
19
1
k=2
19 √
k
k
+
19 √
k
← constant times a sum
k=2
≈ 2(2.5477) + 56.1938 = 61.2892. ← add previously computed values
EXAMPLE 5
Using sum formulas to calculate the value of a sum
3
2
Find the value of the sum 300
k=1 (k − 4k + 2).
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.1
Addition and Accumulation
323
SOLUTION
By writing the sum as three simple sums and using the formulas in Theorem 4.4, we have
300
300
300
300
(k 3 − 4k 2 + 2) =
k3 − 4
k2 + 2
1
k=1
=
k=1
3002 (301)2
k=1
−4
4
= 2, 002, 342, 900.
← properties of sums
k=1
300(301)(601)
+ 2(300)
6
← Theorem 4.4
Of course that was a whole lot easier than expanding the sum and then adding all
300 terms!
EXAMPLE 6
Using a sum formula to calculate the limit of a sum as n → ∞
k2
Find the values of the sum nk=1 3 for n = 3, n = 4, n = 100, and n = 1000. Then
n
investigate what happens as n approaches infinity.
SOLUTION
To find the sums for n = 3 and n = 4, it is easy to expand and calculate the sums directly:
3
k2
k=1
n3
4
k2
k=1
n3
=
12
22
32
1
4
9
14
+ 3 + 3 =
+
+
= ;
33
3
3
27
27
27
27
=
12
22
32
42
1
4
9
16
30
15
+ 3 + 3 + 3 =
+
+
+
=
= .
43
4
4
4
64
64
64
64
64
32
Notice that the denominator of each term in the first sum is constantly equal to n3 and
does not change as k changes. Notice also that the sum from n = 1 to n = 4 cannot be
obtained from the sum from n = 1 to n = 3 just by adding a fourth term; all the terms
in the second sum are different from the terms in the first sum. In fact, the second sum is
actually smaller than the first sum, since
14
27
≈ 0.5185 and
15
32
= 0.46875. Although we are
adding more terms in the second sum, each of those terms is smaller than the terms in the
first sum (since their denominators are larger).
To find the sum for n = 100, we will use properties of sums and one of the sum formulas
from Theorem 4.4:
100 2
k
k=1
n3
=
100
100
1 2
k2
=
k
1003
1003
k=1
← pull out the constant
k=1
1 100(101)(201)
1003
6
2, 030, 100
= 0.33835.
=
6, 000, 000
=
1
1003
← sum formula from Theorem 4.4
Similarly, the sum from k = 1 to k = 1000 is
1000
k=1
1000
1000
k2
1 2
k2
=
=
k
n3
10003
10003
k=1
1
=
10003
k=1
1000(1001)(2001)
6
=
2, 003, 001, 000
≈ 0.33383.
6, 000, 000, 000
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
324
Chapter 4
November 21, 2012
Definite Integrals
Notice that even though we just summed up a thousand terms, the sum is a very small
number. In fact, the more terms we sum up, the smaller that number seems to get! It looks
like if we let n continue to grow, the sum from k = 1 to k = n will approach a number that
is approximately 0.33 or so. We can show that this is indeed the case by taking a limit of
the formula we used to compute the two previous sums:
lim
n→∞
n
k2
k=1
n3
n
1 2
k
n→∞ n3
= lim
← n is constant with respect to k
k=1
1 n(n + 1)(2n + 1)
n→∞ n3
6
(n + 1)(2n + 1)
= lim
n→∞
6n2
2
2n + 3n + 1
2
1
= lim
= = .
6
3
n→∞
6n2
= lim
TEST YOUR
? UNDERSTANDING
← sum formula from Theorem 4.4
← algebra
← take the limit
Considering the stoplight example at the beginning of this section, how can the approx-
imate distances travelled be interpreted as rectangles? How is exact distance travelled
related to the area under a velocity curve?
Can you find an example in which nk=1 (a k b k ) is not equal to the product of nk=1 a k
and nk=1 b k ?
Do two sums in sigma notation have to start and end at the same index value in order
to be combined into one sum? Why or why not?
How can we express the sum in Example 2 by using sigma notation that starts at k = 4
and ends at k = 10?
How could a sum of infinitely many things ever be a small finite number, as happened
in Example 6?
EXERCISES 4.1
Thinking Back
Approximations and limits: Describe in your own words
how the slope of a tangent line can be approximated
by the slope of a nearby secant line. Then describe how
the derivative of a function at a point is defined as a
limit of slopes of secant lines. What is the approximation/limit situation described in this section?
Properties of addition: State the associative law for addition, the commutative law for addition, and the distributive law for multiplication over addition of real
numbers. (You may have to think back to a previous
algebra course.)
Sum and constant-multiple rules: State the sum and
constant-multiple rules for (a) derivatives and (b)
limits.
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The sum formulas in Theorem 4.4 can
be applied only to sums whose starting index value is
k = 1.
(b) True or False:
n k 3 + k 2 + 1
k=0
k+1
(c) True or False:
n k 3 + k 2 + 1
k=1
k+1
(d) True or False:
n
k2
.
k=1
n
k=0
.
n
k=1
n
1
2
+
is equal to
k=1 k
k+1
n
1
2
+
is equal to
k=0 k
k+1
.
n
k=1
1
k+1
n
k+1
CONFIRMING PAGES
k=1
k2
is equal to
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
m √
n √
(e) True or False:
k +
k is equal to
k=0
k=m
n √
k.
k=0
(f) True or False: nk=0 a k = −a 0 − a n + n−1
k=1 a k .
10
2
10
2
(g) True or False:
= k=1 a k .
k=1 a k
n
e x (e x + 1)(2e x + 1)
x 2
.
(h) True or False:
k=1 (e ) =
6
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A sum that would not be suitable for expressing in
sigma notation.
(b) Two different sigma notation expressions of the same
sum.
(c) A sum from k = 1 to k = n that is smaller for n = 10
than it is for n = 5.
3. Explain why it would be difficult to write the sum
1
1
1
1
1
+ +
+
+
in sigma notation.
5
8
11
12
13
4. Use a sentence to describe what the notation
means. (Hint: Start with “The sum of....”)
1 1
+ +
3 4
100 √
k
k=2
2
5. Use a sentence to describe what the notation 87
k=3 k
means. (Hint: Start with “The sum of....”)
6. Consider the general sigma notation nk=m a k . What do
we mean when we say that a k is a function of k?
q
7. Consider the sum i=p bi .
(a) Write out this sum in expanded form (i.e., without
sigma notation).
(b) What is the index of the sum? What is the starting
value? What is the ending value? Which part of the
notation describes the form of each of the terms in
the sum?
(c) Do p and q have to be integers? Can they be negative? What about bi ? What else can you say about p
and q?
k
. Identify the terms a 2 , a 3 ,
8. Consider the sum 5k=2
a4 , and a5 .
1−k
n
9. Consider the sum k=m a k = 9 + 16 + 25 + 36 + 49.
What is a k ? What is m? What is n?
10. Show that
325
4.1
Addition and Accumulation
9
k
k−1
is equal to 10
by writing out
k=4
k+1
k
k=3
the terms in each sum.
8
1
11. Show that
k=0
k2 + 1
is equal to 2
8
k=0
ing out the terms in each sum.
12. Write the sum
1
by writ2k 2 + 2
4
5
6
7
8
+ + +
+
in sigma notation in
7
8
9
10
11
three ways: with a starting value of (a) k = 4, (b) k = 7,
and (c) k = 5.
13. Write the sum 2 +
2
2
2
2
+ +
+
in sigma notation in
4
9
16
25
three ways: with a starting value of (a) k = 1, (b) k = 2,
and (c) k = 0.
√
k into three sums, each in sigma no14. Split the sum 11
k=4
tation, where the first sum has two terms and the last two
sums each have three terms.
n(n + 1)
for the cases (a)
15. Verify that nk=1 k is equal to
2
n = 2, (b) n = 8, and (c) n = 9.
n(n + 1)(2n + 1)
for the cases
16. Verify that nk=1 k 2 is equal to
6
(a) n = 1, (b) n = 5, and (c) n = 10.
17. State algebraic formulas that express the following sums,
where n is a positive integer:
n
n
n
n
(a)
1
(b)
k
(c)
k2
(d)
k3
k=1
k=1
k=1
k=1
18. Explain why terms in the sum in Example 6 with n = 4
are completely different from the terms in the sum when
n = 3. How can the sum from k = 1 to k = 4 be smaller
than the sum from k = 1 to k = 3? What will happen as
n gets larger in this example?
19. Considering the discussion at the end of the stoplight example in the reading, would you expect that the area under the graph of a function f is related to the derivative
f ? Or would you expect that the area under the graph of
a derivative function f is related to the function f ?
20. Consider again the stoplight example from the reading. In
making an approximation for distance travelled, why do
we assume that velocity is constant on small subintervals?
What are some different ways that we could choose which
velocity to use on each subinterval? Illustrate a couple of
these ways with graphs that involve rectangles.
Skills
Write each of the sums in Exercises 21–28 in sigma notation.
Identify m, n, and a k in each problem.
Write out each sum in Exercises 29–34 in expanded form, and
then calculate the value of the sum.
21. 3 + 3 + 3 + 3 + 3 + 3 + 3 + 3
4
5
6
7
8
9
10
11
22.
+ + + + + +
+
3
4
5
6
7
8
9
10
4
5
6
7
23. 3 + +
+
+
8
27
64
125
1
1
1
1
1
1
1
24.
+ +
+
+
+
+
4
9
16
25
36
49
64
25. 5 + 10 + 17 + 26 + 37 + 50 + 65 + 82 + 101
26. 9 + 12 + 15 + 18 + 21 + 24 + 27
27.
1
2
3
n
+ + + ··· +
n
n
n
n
29.
9
k2
30.
k=4
31.
5
k=0
33.
9
k=0
6
k=0
1 2
k
2
1
2
1
3+ k
10
32.
10
2
k+1
ln k
k=3
1
10
34.
4
((2 + k)2 + 1)
k=1
28. −2n − 1n + 0n + 1n + · · · + nn
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
326
Chapter 4
November 21, 2012
Definite Integrals
Find a formula for each of the sums in Exercises 35–40, and
then use these formulas to calculate each sum for n = 100,
n = 500, and n = 1000.
35.
n
(3 − k)
n
(k 3 − 10k 2 + 2)
36.
k=1
k=1
n
(k + 1)2
37.
n
k3 − 1
38.
4
k=3
k=1
n
k3 − 1
39.
n4
n
k2 + k + 1
40.
n3
k=1
k=1
Write each expression in Exercises 41–43 in one sigma notation (with some extra terms added to or subtracted from the
sum, as necessary).
41. 2
100
ak −
k=0
43. 3
25
101
ak
40
1
42.
k=3
k2 + 2
k=2
k=1
24
k−
k=2
25
k
−
39
k=0
In Exercises 44–46, find the sum or quantity without completely expanding or calculating any sums.
10
44. Given 10
k=3 a k = 12 and
k=2 a k = 23, find a 2 .
4
4
45. Given k=1 a k = 7, k=0 b k = 10, and a 0 = 2, find the
value of 4k=0 (2a k + 3b k ).
28
2
46. Given 25
k=0 k = 325 and
k=3 (k − 3) = 14, 910, find the
25 2
value of k=3 (k − 5k + 9).
Determine which of the limit of sums in Exercises 47–52 are
infinite and which are finite. For each limit of sums that is
finite, compute its value.
47.
n
k2 + k + 1
n→∞
n3
lim
48.
k=1
1
k+1
n
(k + 1)2
49. lim
n→∞
n3 − 1
k=1
2
n
k
1
51. lim
·
1+
1
k=0
n→∞
k=1
n
n
lim
n→∞
n
(k 2 + k + 1)
k=1
n
k2 + k + 1
50. lim
n→∞
n2
k=1
52.
lim
n→∞
n
k=1
n4
k3
+n+1
Applications
53. Considering the stoplight example in the reading with
velocity v(t) = −0.22t 2 + 8.8t as shown next at the left,
approximate the distance travelled by dividing the time
interval [0, 40] into eight pieces and assuming constant
velocity on each piece. Interpret this distance in terms of
rectangles on the graph of v(t).
Velocity of car
v(t) = −0.22t 2 + 8.8t
Piecewise approximation of
velocity v(t) of race car
Activity in tuition savings account
v
v
Year
100
88
30
40
2006
2007
2008
1200
60
Earnings $ 10 $ 183 $ 317 $ −1650
40
Increase $ 610 $ 1383 $ 1517 $ −450
20
20
2005
Deposited $ 600 $ 1200 $ 1200 $
80
10
55. The table that follows describes the activity in a college
tuition savings account over four years. Notice that 2008
was a particularly bad year for investing! Let I(t) be the
amount by which your account increased or decreased in
year t, and let B(t) be the balance of your account at the
end of year t.
t
2
4
6
8
10
t
54. Suppose you drive on a racetrack for 10 minutes with
velocity as shown in the graph at the right.
(a) Describe in words the behavior of your race car over
the 10 minutes as shown in the graph.
(b) Find a piecewise-defined formula for your velocity
v(t), in miles per hour, t hours after you start from
rest. (Note that 1 minute is
1
of an hour.)
60
(c) Approximate the distance you travelled over the
10 minutes by using 10 subintervals of 1 minute over
which you assume a constant velocity. Illustrate this
approximation by showing rectangles on the graph
of v(t).
(d) Given that distance travelled is the area under the
velocity graph, use triangles and squares to calculate
the exact distance travelled.
Balance $ 610 $ 1994 $ 3512 $
3061
(a) Describe in your own words how B(t) is the accumulation function for I(t).
(b) Plot a step-function graph of I(t), and describe how
B(t) relates to the area under that graph.
(c) What, if anything, can you say about B(t) when I(t)
is positive? Negative? If you had to guess that one
of these functions was related to the derivative of the
other, which one would it be?
56. Suppose 100 mg of a drug is administered to a patient
each morning in pill form and it is known that after
24 hours the body processes 80% of the drug from such
a pill, leaving 20% of the drug in the body. The amount
of the drug in the body right after the first pill is taken is
A(1) = 100 mg. 24 hours later, after the second pill has
been taken, the amount in the body is A(2) = 100(0.2) +
100 = 120 mg. 48 hours later, the amount in the body
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.1
right after taking the third pill is A(3) = 100(0.2)(0.2) +
100(0.2) + 100 = 124 mg.
Repeated doses of a drug
A
100
80
60
327
Addition and Accumulation
income, and that these transactions continue. Let A(n) be
the accumulated amount of spending, in billions of dollars, that has occurred after n such transactions. For example, A(1) is the amount of spending that has occurred after
the first group of people has spent its money, so A(1) =
$10(0.7) = $7 billion. A(2) is the amount of spending that
has occurred after the first and second groups of people
have spent their money, so A(2) = $10(0.7) + $7(0.7) =
$11.9 billion, as shown in the following graph:
40
Accumulation of tax cut spending
20
1
2
3
4
A
t
25
(a) Explain the calculations for A(1), A(2), and A(3) described in the exercise. Which term in A(3) corresponds to the drug left from the first pill?
(b) Interpret the given graph in the context of this
problem. What do the marked points represent?
(c) Express A(n) in sigma notation.
(d) Calculate the amount of drug in the body after the
4th through 10th pills. Do you notice anything special about A(n) as n gets larger?
57. Suppose the government enacts a $10 billion tax cut
and that the people who save money from this tax cut
will spend 70% of it and save the rest. This generates
$10(0.7) = $7 billion of extra income for other people. Assume these people also spend 70% of their extra
20
15
10
5
1 2 3 4 5 6 7 8
n
(a) Express A(n) in sigma notation.
(b) Calculate A(3), A(4), and A(5).
(c) Estimate the total spending created by this tax cut by
calculating the accumulated spending for at least 10
of these transactions. Interpret your answer in terms
of the given graph.
Proofs
n
58. Give a simple proof that nk=5 (a k + b k ) =
k=5 a k +
n
b
.
k=5 k
59. Give a simple proof that nk=0 3a k = 3 nk=0 a k .
60. Give a simple proof that
if n is a positive integer and c is
any real number, then nk=1 c = cn.
61. Prove part (b) of Theorem 4.4 in the case when n is even: If
n(n + 1)
n is a positive even integer, then nk=1 k =
. (Hint:
Try some examples first, such as n = 6 and n = 10, and think
about how to group the terms to get the sum quickly.)
62. Prove part (b) of Theorem 4.4 in the case when n is odd: If
n(n + 1)
n is a positive odd integer, then nk=1 k =
. (Hint:
2
Use a method similar to the one for the previous exercise, but
take note of what happens with the extra middle term of the
sum.)
2
Thinking Forward
Functions defined by area accumulation: Let f be the function
that is shown here at the left, and define a new function A so
that for every c ≥ 1, A(c) is the area of the region between the
graph of f and the x-axis over the interval [1, c]. For example, A(2) is the area of the shaded region in the graph at the
right.
Use the graph of f to estimate the values of A(1),
A(2), and A(3). (Hint: Consider the grid lines in the graph
shown at the right.)
Describe the intervals on which the function f is positive, negative, increasing, and decreasing. Then describe the intervals on which the function A is positive,
negative, increasing, and decreasing.
From the figures, we can see that f is increasing and
positive on [1, ∞) and A is also increasing and positive on [1, ∞). What would you be able to say about
the area accumulation function A if f were instead
decreasing and positive? Or increasing and negative?
Draw some pictures in your investigation.
Graph of y = f (x)
The shaded area is A(2)
y
y
6
6
5
5
4
4
3
3
2
2
1
1
1
2
3
x
1
2
3
x
Approximating the area under a curve with rectangles: Suppose
you want to find the area between the graph of a positive
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
328
Chapter 4
November 21, 2012
Definite Integrals
function f and the x-axis from x = a to x = b. We can approximate such an area by using a sum of areas of small rectangles
whose heights depend on the height f (x) at various points. For
the problems that follow you should choose rectangles so that
each rectangle has the same width and the top left corner of
each rectangle intersects the graph of f .
4.2
Approximate the area between the graph of f (x) = x 2
and the x-axis from x = 0 to x = 4, by using four rectangles. Include a picture of the rectangles that you are
using.
Sequences of partial sums: In Exercise 57 we saw a function A(n)
that was defined as a sum of n terms, A(n) = nk=1 10(0.7)k .
What happens as n approaches infinity? The sum A(n) is called
a partial sum because it represents part of the sum that accumulates if you let n approach infinity.
Consider the sequence A(1), A(2), A(3), . . ., A(n).
Write out this sequence up to n = 10. What do you
notice?
As n approaches infinity, this sequence of partial
sums could either converge, meaning that the terms
eventually approach some finite limit, or it could
diverge to infinity, meaning that the terms eventually
grow without bound. Which do you think is the case
here, and why?
Approximate the same area as earlier, but this time
with eight rectangles. Is this an over-approximation or
an under-approximation of the exact area under the
graph?
RIEMANN SUMS
Geometric approximation by the process of subdividing, approximating, and adding up
Using rectangles to approximate the area under a curve
Definition and types of Riemann sums in formal mathematical notation
Subdivide, Approximate, and Add Up
As you well know, the formula for finding the area of a circle of radius r is A = π r 2 . In
particular, a circle of radius 2 units has area A = π 22 = 4π. But wait a moment; where
does this area formula come from? Why is it true? Suppose for a moment that we don’t
know the area formula. How could we find, or at least approximate, the area of a circle of
radius 2? The three diagrams that follow suggest an answer.
2
2
2
In the figure at the left, a circle of radius 2 units is shown with a grid where each square
has side length 1 unit and thus area 1 square unit. We need only count up the approximate number of squares to approximate the area of the circle. The circle encloses four full
squares, and 12 partial squares. We will approximate by counting each partial square as half
of a square. This is just one of many approximation methods we could use. This method
produces the approximation
A ≈ 4(1) + 12
1
2
= 10 square units.
As we know, the actual area of the circle is 4π ≈ 12.5664, so the approximation we
found is not very accurate. If instead we use the grid in the second figure shown previously,
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
329
Riemann Sums
we can obtain a better approximation. In this case, the circle covers 32 full squares each
of area
1 2
2
=
1
4
and 28 partial squares (which we will count as half-squares, or
1
8
square
unit each). We now have the approximation
1
4
A ≈ 32
+ 28
1
8
= 11.5 square units.
This approximation is closer to the area we expected, since the squares we used were
smaller. An even better approximation can be obtained by using the rightmost figure shown
earlier, in which there are 164 full squares of area
1
32
1 2
4
=
1
16
and 60 partial squares that
we’ll count as having an area of square units each. This gives us the even better approximation
1
1
A ≈ 164
+ 60
= 12.125 square units.
16
32
We just used a grid to subdivide a circle into a number of smaller pieces, approximated
the area of each of those pieces (by counting each piece as either a full square or a halfsquare), and then added up each of the small approximated areas. This process of “subdividing, approximating, and adding up” is the cornerstone of the definite integral, which
we will introduce in Section 4.3. After we have learned the theory of integrals, we will
be able to prove that the area formula A = πr 2 for a circle of radius r is correct. (See
Section 5.5.)
Before moving on, notice that we could think of each one of our approximations of
the area of a circle as the output of a sort of approximation function whose inputs are the
possible grid sizes. Smaller grid sizes should produce more accurate approximations. As
the grid sizes get smaller, if those approximations somehow stabilize at a real number,
then we will say that the area of the object is the number to which the approximations
stabilize. This is yet another place where our study of limits will pay off! We will make the
ideas in this paragraph more precise in Section 4.3.
Approximating Area with Rectangles
In the rest of this chapter we will be concerned primarily with finding or approximating
the area under a curve, that is, the area enclosed between the graph of some function and
the x-axis on some interval [a, b]. To keep things simple we will begin by restricting our
attention to positive functions. For example, we might be interested in calculating the area
between the graph of the function f and the x-axis from x = 0 to x = 2 as shown here:
Area between f and the x-axis on [0, 2]
y
1
1
2
x
We could use a grid of squares to approximate this area, but it is more efficient to use
a set of rectangles whose heights depend on the height of the function f . For example, we
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
330
Chapter 4
November 21, 2012
Definite Integrals
could subdivide the interval [0, 2] into four smaller intervals of width
1
2
as shown next at
the left. Then we could use the heights of f at the leftmost points of the subintervals to
define four rectangles, as shown in the middle and rightmost figures.
Get heights by using the
left of each subinterval
Subdivide interval [0, 2]
into four subintervals
Create rectangles
to approximate area
y
y
1
y
1
1
1
x
2
1
2
x
1
2
x
1
2
In the third figure, the first rectangle has height f (0) and width . Therefore, the area
1
2
of the first rectangle is f (0)
. The remaining three rectangles have heights f (0.5), f (1),
1
and f (1.5), respectively, all with width . Therefore the sum of the areas of these four rect2
angles is
f (0)
1
2
+ f (0.5)
1
2
+ f (1)
1
2
+ f (1.5)
1
.
2
The function f we have been working with happens to have the unwieldy equation
f (x) =
1
(125x 3 − 325x 2 + 175x + 89).
80
Evaluating f (x) at x = 0, x = 0.5, x = 1, and x = 1.5, we see that the sum of the areas of
the four rectangles is approximately
(1.11)
1
2
+ (1.39)
1
2
+ (0.80)
1
2
+ (0.53)
1
2
≈ 1.915.
The combined area of the rectangles represents a rough approximation of the area under
the graph of f on [0, 2]. Notice that this answer is reasonable, since the region whose area
we are approximating looks to be about half of the size of the square defined by 0 ≤ x ≤ 2
and 0 ≤ y ≤ 2, a square with area 4.
Riemann Sums
We will now develop mathematical notation to formalize the method of using rectangles to
approximate the area under a curve. This formalization will eventually enable us to apply
limits so that we can find the actual area under a curve. Suppose f is a function that is
nonnegative on an interval [a, b]. Consider the area under the graph of f on this interval,
that is, the area of the region bounded above by the graph of f , below by the x-axis, and to
the left and right by the lines x = a and x = b. We will use the subdivide, approximate, and
add up process to define a Riemann sum that will help us to approximate this area.
Subdivide: First, we subdivide the interval [a, b] into n subintervals of equal width x.
b−a
This means that x =
. We will give names to the subdivision points between the
n
subintervals as shown next at the left. Notice that x 0 = a, x 1 = a + x, x 2 = a + 2x, and
so on, until we end at x n = b. In general, the kth subdivision point in the subdivision x k is
equal to a + kx.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
Subdividing [a, b] into n subintervals
a
x1
x
x2
x
x3
x4
...
331
Height and width of kth rectangle
x
b
x0
Riemann Sums
xn
x
f(x k*)
x k1 x k*
xk
Approximate: We now need to define, for each subinterval [x k−1 , x k ], a rectangle that
approximates the area under the graph of f on that subinterval. We need to choose a
height for the rectangle so that the height is somehow related to the height of the graph of
f . In our previous example we chose the leftmost point of each subinterval, but in general
we can choose any point x k∗ in [x k−1 , x k ]. The star in the x k∗ is supposed to suggest that we
are choosing any point we like in that subinterval. Now we can define a rectangle over the
kth subinterval [x k−1 , x k ] such that the height of the rectangle is f (x k∗ ), as shown previously
at the right.
Add up: If we choose n large enough—or equivalently, make x small enough—then
the combined areas of the rectangles we define with the points x k∗ will be a decent approximation for the area under the curve. It now remains only to add up the areas of these
rectangles. The area of the kth rectangle is its height times its width, or f (x k∗ )x, so the total
area under the curve is approximated by the sum
f (x 1∗ )x + f (x 2∗ )x + f (x 3∗ )x + · · · + f (x k∗ )x + · · · + f (x∗n )x.
This sum can be expressed very compactly in sigma notation:
n
f (x k∗ )x.
k=1
Note that to find this sum, we followed a “subdivide, approximate, and add” process. Sums of this form are called Riemann sums, named for the prolific mathematician
Bernhard Riemann, who developed the notion. The following definition summarizes the
notation that we have developed:
DEFINITION 4.5
Riemann Sums
A Riemann sum for a function f on an interval [a, b] is a sum of the form
n
where x =
b−a
,
n
x k = a + kx,
k=1
and x k∗
f (x k∗ ) x,
is some point in the interval [x k−1 , x k ].
As we have seen, a Riemann sum for a function f on an interval [a, b] approximates the
area under the graph of f between x = a and x = b. If f is continuous on [a, b], then
as the number of rectangles n approaches infinity, the approximation will get better and
better and approach the actual area under the graph of f on [a, b]. We will make this idea
mathematically precise in Section 4.3.
It is possible to consider Riemann sums for which the interval [a, b] is partitioned into
n subintervals by points a = x 0 , x 1 , x 2 , . . . , x n−1 , x n = b that are not equal distances apart.
In such subdivisions we would have a different width (x)k for each subinterval [x k−1 , x k ].
This more general theory of Riemann sums defined with arbitrary partitions is important
in later mathematics courses, but is not needed for our study of calculus here.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
332
Chapter 4
November 21, 2012
Definite Integrals
Signed Area
When the graph of a function lies below the x-axis, we can still use rectangles to approximate area. For example, consider the following Riemann sum with n = 4 rectangles that
approximates the area of the region between the graph of f (x) = x 2 − 1 and the x-axis on
[0, 1]:
A Riemann sum with n = 4 for
f (x) = x 2 − 1 on [0, 1]
y
0.25
0.50
0.75
1.00
x
1
Note that the “height” of each of the rectangles in this Riemann sum is negative. For
example, the first rectangle has a height of f (0) = 02 − 1 = −1 and is counted as having an
“area” of −0.25. The left sum in this case is the sum of the areas of these four “negative”
rectangles, −0.25 − 0.234375 − 0.1875 − 0.109375 = −0.78125.
Since, by construction, Riemann sums automatically count the area between a negative
function and the x-axis as a negative number and the area between a positive function and
the x-axis as a positive number, we will say that Riemann sums measure signed area, also
known as net area. For example, if f is the following function graphed on [0, 5], then the
right sum, with 25 rectangles, will have some rectangles counting area positively and some
rectangles counting area negatively, depending on whether f is above or below the x-axis:
Area is counted positively on [0, 2] and [4, 5], and negatively on [2, 4]
y
10
5
1
2
3
4
5
x
5
Types of Riemann Sums
Depending on how we choose the point x k∗ in each subinterval [x k−1 , x k ], we get different
types of Riemann sums. For example, we could choose each x k∗ to be the leftmost point
of the kth subinterval, as shown in the first graph that follows; this is called a left sum.
Alternatively, we could construct a right sum by choosing x k∗ to be the rightmost point of
the kth subinterval, or a midpoint sum by choosing x k∗ to be the midpoint, as shown in the
following middle and rightmost graphs, respectively:
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
333
Riemann Sums
Leftmost point of each
subinterval determines heights
Rightmost point of each
subinterval determines heights
Midpoint of each
subinterval determines heights
y
y
y
1
1
1
2
1
x
1
x
2
1
2
x
Algebraically, these three ways of choosing x k∗ from each subinterval [x k−1 , x k ] yield the
following three types of Riemann sums:
DEFINITION 4.6
Left, Right, and Midpoint Sums
Suppose f is a function defined on the interval [a, b]. Given a positive integer n, let
b−a
x =
and x k = a + kx. Then
n
(a) The n-rectangle left sum for f on [a, b] is nk=1 f (x k−1 ) x.
(b) The n-rectangle right sum for f on [a, b] is
n
k=1 f (x k ) x.
(c) The n-rectangle midpoint sum for f on [a, b] is
n
k=1 f
xk−1 + x k
2
x.
Note that given any interval [a, b] and number n of rectangles, we can write x and x k
in terms of a, b, and n. In practice, we will always need to use the explicit expressions
b−a
x =
and x k = a + kx (as well as using the definition of the function f ) when
n
evaluating a Riemann sum. For example, the right sum expressed earlier is equal to
n
f a+k
k=1
b−a
n
b−a
.
n
We can also set up Riemann sums that we can guarantee will be over-approximations
or under-approximations of the actual area under the graph of a continuous function. The
first figure that follows shows the upper sum, where each x k∗ is chosen so that f (x k∗ ) is the
maximum value of f on [x k−1 , x k ]. The upper sum is always greater than or equal to the
actual signed area. Similarly, in the lower sum shown in the second figure we choose x k∗
so that f (x k∗ ) is the minimum value of f on [x k−1 , x k ]. The lower sum is always less than or
equal to the actual signed area.
Maximum value of f on each
subinterval determines heights
Minimum value of f on each
subinterval determines heights
y
y
1
1
1
2
x
1
CONFIRMING PAGES
2
x
18:50
TKmaster2010
WHF00153/FREE087-Taalman
334
Chapter 4
November 21, 2012
Definite Integrals
Algebraically, we can express these two types of Riemann sums as follows:
DEFINITION 4.7
Upper and Lower Sums
Suppose f is a function that is continuous on the interval [a, b]. Given a positive integer n,
b−a
and x k = a + kx. Then
let x =
n
(a) The n-rectangle upper sum for f on [a, b] is nk=1 f (Mk ) x, where each Mk is
chosen so that f (Mk ) is the maximum value of f on [x k−1 , x k ].
(b) The n-rectangle lower sum for f on [a, b] is nk=1 f (mk ) x, where each mk is chosen
so that f (mk ) is the minimum value of f on [x k−1 , x k ].
The reason we require that f be continuous on [a, b] is that this is necessary in order for
the Extreme Value Theorem to guarantee that f attains maximum and minimum values on
each subinterval.
The upper sum and the lower sum can be more complicated to calculate than the left,
right, or midpoint sums, but they can be useful if we wish to not only approximate the
area under a curve, but get a bound on how much error is involved in our approximation.
For example, if an upper sum approximation for the area under a curve was calculated to
be 4.4 and a lower sum approximation for the same area was calculated to be 4.1, then
we would know that the actual area was in the interval [4.1, 4.4]. In particular, this would
mean that both the upper sum and the lower sum were within 0.3 square unit of the actual
area under the curve.
We also don’t necessarily have to use rectangles to approximate the area under a curve.
For example, we could use trapezoids instead, as shown here:
Trapezoids on each subinterval
approximate area under a curve
y
1
1
x
2
The trapezoid sum uses a trapezoid with width x, left height f (x k−1 ), and right height
f (x k ) to approximate the slice of area in the subinterval [x k−1 , x k ]. Recall that the area of a
trapezoid with base b and heights h1 and h2 is
the following algebraic definition:
DEFINITION 4.8
h1 + h2
2
b. This formula for the area suggests
Trapezoid Sums
Suppose f is a function defined on the interval [a, b]. Given a positive integer n, let
b−a
x =
and x k = a + kx. Then
n
The n-rectangle trapezoid sum for f on [a, b] is
n
f (x
k=1
k−1 )+f (x k )
2
x.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
Riemann Sums
335
TECHNICAL POINT To be honest, in the form of Definition 4.8, the trapezoid sum is not
technically a Riemann sum for f ; recall that the terms in a Riemann sum must be of the
form f (x k∗ )x for some x k∗ ∈ [x k−1 , x k ]. However, if f is a continuous function, then in each
subinterval we can use the Intermediate Value Theorem to find some x k∗ in for which f (x k∗ )
is equal to the average of f (x k−1 ) and f (x k ). This means that when f is continuous, the
trapezoid sum really is a Riemann sum “in disguise.”
Examples and Explorations
EXAMPLE 1
Approximating area with a right sum
Approximate the area between the graph of f (x) = x 2 − 2x + 2 and the x-axis from x = 1
to x = 3, using a right sum with four rectangles.
SOLUTION
We begin by subdividing the interval [1, 3] into four subintervals: [1, 1.5], [1.5, 2], [2, 2.5],
1
and [2.5, 3], each of width . The rightmost points of these four subintervals are, respec2
tively, 1.5, 2, 2.5, and 3. The values of f at these points will be the heights of our four
rectangles, as shown here:
Right sum with four rectangles
y
6
5
4
3
2
1
1
2
3
x
It is now a simple matter to add up the areas of the four rectangles:
1
1
+ f (2)
+ f (2.5)
2
2
1
1
= (1.25)
+ (2)
+ (3.25)
2
2
1
1
+ f (3)
2
2
1
1
+ (5)
2
2
Area ≈ f (1.5)
← plug into f (x) = x 2 − 2x + 2
= 5.75.
EXAMPLE 2
Understanding the notation used in Riemann sums
Looking back at Example 1, identify each of the following:
a, b;
n;
x;
x 0 , x 1 , x 2 , x 3 , x4 ;
x 1∗ , x 2∗ , x 3∗ , x 4∗ .
SOLUTION
1
2
We have a = 1, b = 3, n = 4, and x = . The subdivision points are x 0 = a = 1, x 1 = 1.5,
x 2 = 2, x 3 = 2.5, and x4 = b = 3. The chosen points in each subinterval [x k−1 , x k ] are
taken to be the rightmost points: x 1∗ = 1.5, x 2∗ = 2, x 3∗ = 2.5, and x 4∗ = 3. Notice that since
this is a right sum, each x k∗ is just equal to x k .
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
336
Chapter 4
EXAMPLE 3
November 21, 2012
Definite Integrals
Expressing a right sum in sigma notation
Consider again the area between the graph of f (x) = x 2 − 2x + 2 and the x-axis from x = 1
to x = 3. Express the n-rectangle right-sum approximation in sigma notation. Simplify until
k and n are the only variables that appear in your expression. Then use your expression to
calculate the right-sum approximations with n = 4 and n = 8 rectangles.
SOLUTION
We first subdivide [1, 3] into n subintervals, each of width x =
subdivision points x k are of the form x k = a + kx = 1 +
2k
.
n
b−a
n
=
3−1
n
=
2
.
n
The
Since we are constructing a
right sum, in each subinterval [x k−1 , x k ] we will choose x k∗ to be the rightmost point x k . The
kth rectangle in our approximation will have height f (x k∗ ) and width x. Using the notation
we have developed, we find that the sum of the areas of the n rectangles is
n
n
2k
2
f (x k∗ )x =
f 1+
k=1
=
k=1
n
k=1
n
1+
n
2k 2
2k
−2 1+
n
n
+2
2
n
=
n
1+
k=1
4k 2
n2
2
.
n
Substituting n = 4 into this expression and then writing out the sum gives us the same
four-rectangle right-sum approximation that we found in Example 1:
4
4k 2
2
4(1)2
1
4(2)2
1
= 1+
+ 1+
1+ 2
4
4
16
2
16
2
k=1
4(3)2
1
4(4)2
1
+ 1+
+ 1+
16
1
= (1.25)
2
2
1
+ (2)
2
16
1
+ (3.25)
2
2
+ (5)
1
2
= 5.75.
Similarly, we can substitute n = 8 and then expand the sum to calculate the eight-rectangle
right-sum approximation:
8
4k 2
2
4(1)2
1
4(2)2
1
4(3)2
1
= 1+
+ 1+
+ 1+
1+ 2
(8)
8
64
4
64
4
64
4
k=1
4(4)2
1
4(5)2
1
4(6)2
1
+ 1+
+ 1+
+ 1+
64
4
64
4
64
4
2
2
4(7)
1
4(8)
1
+ 1+
+ 1+
64
4
64
4
1
= (1.0625)
4
1
1
1
+ (1.25)
+ (1.5625)
+ (2)
4
4
4
1
1
1
1
+ (2.5625)
+ (3.25)
+ (4.0625)
+ (5)
4
4
4
4
= 5.1875.
Notice that sigma notation does not in general make it easier to calculate the areas of the
rectangles in our approximations. However, the general n-rectangle sigma notation for a
Riemann sum will be very important when we discuss the theory of definite integrals in
the next section. It would also be very useful if you wanted to write a computer program to
approximate the area under a curve, especially with a large number of rectangles.
EXAMPLE 4
Using the left, midpoint, and trapezoid sums to approximate area
Approximate the area between the graph of f (x) = x 2 − 2x + 2 and the x-axis on [1, 3],
using n = 4 subintervals with (a) the left sum; (b) the midpoint sum; and (c) the trapezoid
sum.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
337
Riemann Sums
SOLUTION
(a) As in Example 1, we have x =
1
2
k
2
and x k = 1 + kx = 1 + . For a left sum we
choose x k∗ to be the leftmost point in the interval [x k−1 , x k ], and therefore we choose
x k∗ = x k−1 = 1 +
4
f 1+
k=1
k−1
.
2
k−1
2
The left sum with four rectangles is therefore equal to
1
2
1
2
= f (1)
+ f (1.5)
1
2
1
2
+ f (2)
+ f (2.5)
1
2
= 3.75.
(b) For a midpoint sum we choose x k∗ to be the midpoint in the interval [x k−1 , x k ], and
therefore we choose x k∗ =
x k−1 + x k
2
=
(1 +
k−1
) + (1 + 2k )
2
2
=
3 + 2k
.
4
The midpoint sum with
four rectangles is therefore equal to
4
k=1
f
3 + 2k
4
1
2
= f (1.25)
1
2
1
2
+ f (1.75)
1
2
+ f (2.25)
+ f (2.75)
1
2
= 4.625.
(c) For the trapezoid sum we do something completely different and use the heights at
k−1
k
and x k = 1 + to calculate the area of the kth trapezoid. The
both x k−1 = 1 +
2
2
trapezoid sum with four trapezoids is equal to
4 f 1 +
k−1
2
+f 1+
k
2
f (1) + f (1.5) 1
f (1.5) + f (2) 1
+
2
2
2
2
k=1
f (2) + f (2.5) 1
f (2.5) + f (3) 1
+
+
= 4.75.
2
2
2
2
The three figures that follow show the three area approximations we just calculated.
Which one do you think is the most accurate, and why? (See Exercise 5.)
1
2
2
Left sum
Midpoint sum
y
y
6
6
6
5
5
5
4
4
4
3
3
3
2
2
2
1
1
TEST YOUR
Trapezoid sum
y
1
? UNDERSTANDING
=
2
3
x
1
1
2
3
x
1
2
3
x
Recalling the discussion of the circle at the start of this section, can you think of a dif-
ferent way of approximating the area of the circle?
The Riemann sums presented in this section are not the only sums we could use to
approximate the area under a curve. For example, we could use a “one-third” sum,
which would define x k∗ to be the point one-third of the way across the interval [x k−1 , x k ].
How would you write x k∗ in terms of x k−1 and x k for the one-third sum? What other
methods for choosing x k∗ can you think of?
Both the midpoint sum and the trapezoid sum use an average somewhere in their for-
mula. What is the average that is being taken in each case?
Can you label x 2 , x 2∗ , and f (x 2∗ ) in each of the Riemann sum figures in this section?
In general, the left sum will always underestimate the area under an increasing function
and overestimate the area under a decreasing function; why? What about the right
sum?
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
338
Chapter 4
November 21, 2012
Definite Integrals
EXERCISES 4.2
Thinking Back
Simple area formulas: Give formulas for the areas of each of
the following geometric figures.
A right triangle with legs of lengths a and b
A triangle with base b and altitude h
A circle of radius r
A rectangle with sides of lengths w and l
A semicircle of radius r
A trapezoid with width w and heights h1 and h2
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The n-rectangle lower sum for f on [a, b]
is always equal to the corresponding right sum.
(b) True or False: If f is positive and increasing on [a, b],
then any left sum for f on [a, b] will be an underapproximation.
(c) True or False: If f (a) > f (b), then any right-sum
approximation for f on [a, b] will be an underapproximation.
(d) True or False: A midpoint sum is always a better
approximation than a left sum.
(e) True or False: If f is positive and concave up on all of
[a, b], then any left-sum approximation for f on [a, b]
will be an under-approximation.
(f) True or False: If f is positive and concave up on all of
[a, b], then any trapezoid sum approximation for f on
[a, b] will be an over-approximation.
(g) True or False: An upper sum approximation for f on
[a, b] can never be an under-approximation.
(h) True or False: For every function f on [a, b], the left
sum is always a better approximation with 10 rectangles than with 5 rectangles.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A graph of a function f on an interval [a, b] for which
the left sum with four rectangles finds the exact area
under the graph of f from x = a to x = b.
(b) A graph of a function f on an interval [a, b] for which
all trapezoid sums (regardless of the size of n) will find
the exact area under the graph of f from x = a to x = b.
(c) A graph of a function f on an interval [a, b] for
which the upper sum with four rectangles is a much
better approximation than the lower sum with four
rectangles.
3. In the reading we used objects whose area we knew
(squares) to approximate the area of a more complicated
object (a circle). The same kind of technique can be used
to approximate the area of the blob pictured here.
(a) Given that each of the squares in the grid has a side
length of
1
square unit, approximate the area of the
2
blob.
(b) How could you get a better approximation?
4. Consider the area between the graph of a function f and
the x-axis from x = 0 to x = 2, as shown in each of the
two figures that follow.
(a) Use the grid on the left and whatever method you
like to approximate this area. Then use the grid on the
right and the same method to make another approximation. Which approximation is likely more accurate,
and why?
(b) Use the grid at the left to get an upper bound on the
area of the region. In other words, make an approximation that you know is greater than the actual area.
Repeat for the grid at the right.
(c) Use the grid at the left to get a lower bound on
the area of the region. In other words, make an approximation that you know is less than the actual area.
Repeat for the grid at the right.
(d) Use your answers to parts (a)–(c) to come up with
your best possible guess for the actual area under the
curve.
y
y
1
1
1
x
CONFIRMING PAGES
1
x
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
14
≈ 4.6667. Given
x 2 − 2x + 2 on [1, 3] is equal to
3
this information, which of the approximations in
Examples 1 and 4 was the most accurate? Was that
what you expected?
6. Do you think that one type of Riemann sum (right, left,
midpoint, upper, lower, trapezoid) is usually more accurate than the others? Why or why not?
7. Suppose that the left-sum approximation with eight rectangles for the area between the graph of a function f and
the x-axis from x = a to x = b is equal to 8.2 and that the
corresponding right sum approximation is equal to 7.5.
(a) What is the corresponding trapezoid sum approximation for this area?
(b) Is the corresponding midpoint sum for this area necessarily between 7.5 and 8.2? If so, explain why. If not,
sketch an example of a function f on an interval [a, b]
whose midpoint sum is not between the left sum and
the right sum.
(c) What can you say about the corresponding upper
sum for this area? The corresponding lower sum?
(d) Is it necessarily true that f is decreasing on the entire interval [a, b]? If so, explain why. If not, sketch an
counterexample in which the left sum is greater than
the right sum but f is not decreasing on all of [a, b].
(e) Could the function f be increasing on the entire interval [a, b]? If not, explain why not. If so, sketch a possible example in which the left sum is greater than the
right sum and f is increasing on all of [a, b].
8. In the reading we mentioned that the trapezoid sum is
the average of the left sum and the right sum. Use the
solutions of Examples 1 and 4 to show that for f (x) =
x 2 − 2x + 2, [a, b] = [1, 3], and n = 4, the trapezoid sum
is indeed the average of the left sum and the right sum.
9. Explain why the upper sum approximation for the area
between the graph of a function f and the x-axis on [a, b]
must always be larger than or equal to any other type of
Riemann sum approximation with the same number n of
rectangles.
10. Consider the area between the graph of a positive function f and the x-axis on an interval [a, b]. Explain why the
11. Suppose you wanted to calculate the upper sum approximation for the area between the graph of f (x) = (x − 1)2
and the x-axis from x = 0 to x = 2. List all of the values
M k used for (a) n = 2 rectangles, (b) n = 3 rectangles, and
(c) n = 4 rectangles. Sketch graphs of your rectangles to
illustrate your answers.
12. Repeat Exercise 11, using the lower sum approximation
and the values mk .
13. Suppose v(t) is the velocity of a particle moving on a
straight path, where v is measured in meters per second
and t is measured in seconds. The particle starts moving
at time t0 and moves for t seconds.
(a) What are the units of v(t0 )t?
(b) Geometrically, what does v(t0 )t represent?
(c) What do these questions have to do with this section?
14. Explain in no more than three sentences how we can
approximate the derivative of a function f at a point c if
we know the graph of f . Then, in no more than three
additional sentences, discuss how the method for approximating area is similar to, and how is it different from,
approximating the derivative. (Both descriptions should
involve multiple approximations, each better than the
last.)
k
(0.25) can’t be a
15. Explain why the sum 20
k=1 f − 3 +
2
right sum for f on [a, b] = [−3, 2].
16. Explain why the sum 100
k=1 f (2 + 0.1(k − 1))(0.1) can’t be
a left sum for f on [a, b] = [2, 5].
Each of the sums in Exercises 17–20 approximates the area between the graph of some function f and the x-axis from x = a
to x = b. Do some “reverse engineering” to determine the
type of approximation (left sum, midpoint sum, etc.) and identify f (x), a, b, n, x, and x k . Then sketch the approximation described.
2
4
k
1
1+
17.
2
k=1
18.
2
ln 2 +
k=0
2
k
3
1
3
100
sin(0.05(k − 1)) + sin(0.05k)
(0.05)
19.
2
k=1
9
1 + k −3 1 + 1 + 3k
1
20.
3
2
k=1
Skills
Your calculator should be able to approximate the area between a graph and the x-axis. Determine how to do this on
your particular calculator, and then, in Exercises 21–26, use the
method to approximate the signed area between the graph of
each function f and the x-axis on the given interval [a, b].
339
upper sum approximation for this area with n = 8 boxes
must be smaller than or equal to the upper sum approximation with n = 4 boxes. It may help to sketch some
examples.
5. In Examples 1 and 4 we found four different approximations for the area between the graph of f (x) = x 2 − 2x + 2
and the x-axis on [1, 3].
(a) Based on the pictures of these approximations given
in the reading, which are over-approximations?
Which are under-approximations? Which approximation looks like it might be closest to the actual area
under the curve?
(b) The actual area (which we won’t know how to calculate until Section 4.5) under the graph of f (x) =
Riemann Sums
21. f (x) =
√
x − 1,
[a, b] = [2, 3]
22. f (x) = x 2 , [a, b] = [0, 3]
23. f (x) = e x , [a, b] = [1, 4]
24. f (x) = sin x, [a, b] = [0, π ]
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
340
Chapter 4
25. f (x) = (x − 2)2 + 1,
November 21, 2012
Definite Integrals
[a, b] = [1, 3]
26. f (x) = 1 − 2 x , [a, b] = [−3, 1]
For each function f and interval [a, b] in Exercises 27–33, use
the given approximation method to approximate the signed
area between the graph of f and the x-axis on [a, b]. Determine whether each of your approximations is likely to be an
over-approximation or an under-approximation of the actual
area.
27. f (x) = x 2 , [a, b] = [0, 3], left sum with
(a) n = 3
(b) n = 6
28. f (x) = sin x, [a, b] = [0, π ], n = 3, with
(a) trapezoid sum
(b) upper sum
√
29. f (x) = x − 1, [a, b] = [2, 3], n = 4, with
(b) right sum
(a) left sum
30. f (x) = 1 − 2 x , [a, b] = [−3, 1], n = 8, with
(b) right sum
(a) left sum
For each function f and interval [a, b] in Exercises 34–38, it is
possible to find the exact signed area between the graph of f
and the x-axis on [a, b] geometrically by using the areas of circles, triangles, and rectangles. Find this exact area, and then
calculate the left, right, midpoint, upper, lower, and trapezoid
sums with n = 4. Which approximation rule is most accurate?
34. f (x) = 5, [a, b] = [−2, 2]
35. f (x) = 3x + 1, [a, b] = [3, 5]
36. f (x) = 4 − x, [a, b] = [0, 6]
√
37. f (x) = 1 − x 2 , [a, b] = [−1, 1]
√
38. f (x) = 3 + 4 − x 2 , [a, b] = [−2, 2]
In Exercises 39–44, write out the sigma notation for the
Riemann sum described in such a way that the only letter
which appears in the general term of the sum is k. Don’t calculate the value of the sum; just write it down in sigma notation.
√
39. f (x) = x − 1, [a, b] = [2, 3], right sum, n = 4.
31. f (x) = e x , [a, b] = [1, 4], n = 6, with
(a) midpoint sum
(b) trapezoid sum
40. f (x) = x 2 , [a, b] = [0, 3], left sum, n = 3.
32. f (x) = 9 − x , [a, b] = [0, 5], n = 5, with
(a) midpoint sum
(b) lower sum
42. f (x) = ln x, [a, b] = [2, 5], left sum, n = 100.
2
33. f (x) = (x − 2)2 + 1, [a, b] = [1, 3], lower sum with
(a) n = 2
(b) n = 3
(c) n = 4
41. f (x) = e x , [a, b] = [1, 4], midpoint sum, n = 6.
43. f (x) = sin x, [a, b] = [0, π ], trapezoid sum, n = 4.
√
44. f (x) = 1 − x 2 , [a, b] = [−1, 1], midpoint sum, n = 20.
Applications
45. Suppose that, as in Section 4.1, you drive in a car for
40 seconds with velocity v(t) = −0.22t 2 + 8.8t feet per
second, as shown in the graph that follows. If your total
distance travelled is equal to the area under the velocity
curve on [0, 40], then find lower and upper bounds for
your distance travelled by using
measured in degrees. The casserole cools by changing at
a rate of T (t) = −15e−0.5t degrees per minute.
Rate of change of casserole temperature
T (t) = −15e−0.5t
T
(a) the lower sum with n = 4 rectangles;
(b) the upper sum with n = 4 rectangles.
1
2
3
4
5
t
5
Velocity of car
v(t) = −0.22t 2 + 8.8t
10
v
15
88
10
20
30
40
t
46. Repeat Exercise 45, using
(a) the midpoint sum with n = 4 rectangles;
(b) the trapezoid sum with n = 4 trapezoids.
If the exact distance travelled is just over 2,346 feet, then
which of these approximations is the most accurate?
47. Dad’s casserole surprise is hot out of the oven, and its
temperature after t minutes is given by the function T(t),
(a) Discuss what the graph of T (t) says about the behavior of the casserole’s temperature after the casserole
is taken out of the oven.
(b) Just as we can approximate distance travelled by
approximating the area under the corresponding
velocity curve, we can approximate the change in
temperature T of the casserole by approximating the
area under the graph of T . Why?
(c) Estimate the change in temperature of the casserole
over the first 5 minutes it is out of the oven, using any
Riemann sum you like, with n = 10.
48. The National Oceanic and Atmospheric Administration
tabulates flow data from many American rivers. From
this data they compute and plot median annual flows.
The flows are given by functions whose closed form is
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.2
not known, but for which we can read off values for any
day of the year we like. The following table describes the
flow f (t) in cubic feet per second for Idaho’s Lochsa River,
t days after January 1:
t
0
60 120 180 240 300 360
f (t) 700 1000 6300 4000 500 650 700
(a) Keeping in mind that the data are periodic with
period 365, compute the left Riemann sum for the
function that these data have sampled.
(b) What is the relation between the numbers you
computed and the total amount of water that
flows down the Lochsa annually? Estimate the total
amount of water that flows down the Lochsa each
year.
(c) Most of the flow down the river takes place from April
to July. We can get a better idea of the total flow if we
add a few data points. Recompute the left Riemann
sum, adding the data points (90, 2100), (150, 11000),
and (210, 1000).
341
Riemann Sums
(d) What would you need to do to get an even better
estimate of the total flow?
49. To approximate the flow f (t) of the Lochsa River in its
flood stage, we can use a function of the form
(t − 90)π
2
−
,
g(t) = c1 + c2 sin
π
105
where the coefficients c1 and c2 are found by evaluating
the following two integrals:
195
1
f (t) dt,
c1 =
105
90
1
c2 =
9.95
195
f (t) sin
90
(t − 90)π
105
−
2
π
dt.
(a) Use the data points (t, f (t)) = (90, 2100), (120, 6300),
(150, 11000), and (180, 4000), and left Riemann sums
to approximate the values of the integrals for c1
and c2 .
(b) Now that you have found c1 and c2 , plot the resulting
function g(t) against the data points from Exercise 48.
Proofs
50. Use Definition 4.7 to prove that for any function f and
interval [a, b], the upper sum with n rectangles is greater
than or equal to the lower sum with n rectangles.
51. Use Definition 4.6 and the definition of increasing to prove
that if a function f is positive and increasing on [a, b], then
the left sum with n rectangles is less than or equal to the
right sum with n rectangles.
52. Use Definitions 4.6 and 4.8 to prove that for any function
f and interval [a, b], the trapezoid sum with n trapezoids
is always the average of the left sum and the right sum
with n rectangles.
53. Use Definition 4.8 to prove that if a function f is positive
and concave up on [a, b], then the trapezoid sum with
n trapezoids is an always an over-approximation for the
actual area.
Thinking Forward
Approximating the length of a curve: Suppose you want to
calculate the driving distance between New York City and
Dallas, Texas.
Print out a highway map of the United States, and
highlight a route, snaking along with the paths of the
major highways.
How could we use the method of “subdivide, approximate, and add up” to approximate this driving distance?
Make an actual approximation of the driving distance
from New York City to Dallas, Texas, using the route
and method you just described.
How accurate do you think your approximation is? Is
it an over-approximation or an under-approximation?
Limits of Riemann sums: In the reading we saw that the
area between the graph of f (x) = x 2 − 2x + 2 and the
x-axis on [1, 3]could be approximated with the right sum
n
4k 2
2
. Let A(n) be equal to this n-rectangle
k=1 1 + 2
n
n
right-sum approximation. The following table describes
various values of A(n):
n
A(n)
10
5.08
100
4.7068
1000
4.6707
10, 000
4.66707
Describe the meaning of the entries in this table, and
verify that the entry for A(10) is correct.
Use the table to make a graph of A(n), and discuss
what happens to this graph as n approaches infinity.
What does your graph tell you about the right-sum
approximations of the area under the graph of f as n
approaches infinity?
Approximations and error: In Section 4.5 we will see that
definite integrals can be computed by taking differences of
antiderivatives; in particular, the Fundamental Theorem of
Calculus will reveal that if f is continuous on [a, b], then
b
a f (x) dx = F(b) − F(a), where F is any antiderivative of f .
Armed with this fact, we can check the exact error of Riemann
sum approximations for integrals of functions that we can antidifferentiate.
Use the given antiderivative fact to find the exact
4 1
1
value of 1 dx. (Hint: What is an antiderivative of ?
x
CONFIRMING PAGES
x
18:50
TKmaster2010
WHF00153/FREE087-Taalman
342
Chapter 4
November 21, 2012
Definite Integrals
In other words, what is a function F whose derivative is
By trial and error with smaller values of n, find
the smallest value of n for which a right sum will
4 1
approximate 1 dx to within 0.25.
Repeat the preceding steps, but with the midpoint
sum in place of the right sum.
1
x
f (x) = ?)
What is the actual error that results from a right-sum
4 1
approximation with n = 4 for 1 dx?
x
x
4.3
DEFINITE INTEGRALS
The exact area under the graph of a function f on [a,b]
Defining the definite integral as a limit of Riemann sums
Properties of the definite integral
Defining the Area Under a Curve
We have seen how to approximate areas with Riemann sums, but how do we find the exact
area between the graph of a function f and the x-axis on an interval [a, b]? Consider the area
of the region bounded by the curve f (x) = x 2 , the x-axis, and the line x = 1 as shown here:
Area under f (x) = x 2 on [0, 1]
y
1
1
x
We don’t yet have a rigorous mathematical method for finding this area exactly. Using
Riemann sums, we can get approximations for the area. For example, the three figures that
follow illustrate upper-sum approximations using n = 10, n = 50, and n = 100. Notice that
in this case, since f (x) = x 2 is increasing on [0, 1], the upper sum happens to be the same
as the right sum.
Upper sum with n = 10
Upper sum with n = 50
y
Upper sum with n = 100
y
y
1.00
1.00
1.00
0.75
0.75
0.75
0.50
0.50
0.50
0.25
0.25
0.25
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
As the number n of rectangles gets larger, the upper sum approximation gets closer to the
area under f (x) = x 2 on [0, 1]. The same thing happens if we consider any other Riemann
sum, say, the lower sum, as pictured here for n = 10, n = 50, and n = 100:
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
Lower sum with n = 10
Lower sum with n = 50
y
Lower sum with n = 100
y
y
1.00
1.00
1.00
0.75
0.75
0.75
0.50
0.50
0.50
0.25
0.25
0.25
0.25
0.50
0.75
1.00
x
343
Definite Integrals
0.25
0.50
0.75
1.00
x
0.25
0.50
0.75
1.00
x
If a sequence of Riemann sum approximations converges to a real number when we
take the limit as n approaches infinity, then we call this number the “area” under the graph.
Notice that up until this point we did not actually have a truly rigorous definition for even
the concept of such an area! The following definition describes this limit and the notation
we will use to refer to it.
DEFINITION 4.9
The Definite Integral of a Function on an Interval
Let f be a function defined on an interval [a, b]. The definite integral of f from x = a to
x = b is defined to be the number
b
n
f (x) dx = lim
f (x k∗ ) x,
n→∞
a
if this limit exists, where x =
[x k−1 , x k ].
b−a
,
n
k=1
x k = a + k x, and x k∗ is any choice of point in
The “ ” symbol should remind you of the letter “S” for “Sum.” A Riemann sum describes
a discrete sum of areas of rectangles, while in a loose sense an integral represents a continuous sum of the areas of infinitely many rectangles, each of which is infinitely thin. The a
and b below and above the integral symbol are called the limits of integration, where here
“limits” is used in the sense of “ends,” not in the sense of limits of functions. It may help
you remember this definition if you think of it in the following way: As n → ∞, the finite
sum
of n things becomes an integral
that accumulates everything from x = a to
x = b. The discrete list of values f (x k∗ ) becomes the continuous function f (x), and the small
change x becomes an “infinitesimal” change dx.
CAUTION
The “dx” portion of the definite integral is not optional; it must always be included with
the notation of the definite integral. Just as
d
(
dx
is how we denote “the derivative of
)
,” the notation
b
dx
a
is how we will denote “the definite integral of
on [a, b],” which represents “the
signed area between the graph of
and the x-axis on [a, b].” Note that in both
cases there is the presence of a “dx” that represents the result of a limiting process
on a change x in the independent variable with respect to which the differentiation or
integration/accumulation/area process is taking place.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
344
Chapter 4
November 21, 2012
Definite Integrals
Consider again the example with f (x) = x 2 and [a, b] = [0, 1]. For any given n, we have
x =
b−a
1
=
n
n
k
n
and x k = a + kx = .
k
We will use the right sum and choose each x k∗ to be the rightmost point x k = of the
n
subinterval [x k−1 , x k ]. When we take the limit as the number n of subdivisions approaches
infinity, we obtain
1
n
k 2 1
.
x 2 dx = lim
n→∞
0
k=1
n
n
If, for a given function f and interval [a, b], the limit defining the integral exists for
any choice of points x k∗ , then we say that f is integrable on [a, b]. It turns out that
every continuous function is integrable on [a, b]. (The proof of this theorem is beyond the
scope of the text and will not be presented here.) However, some discontinuous functions
are still integrable; for example, functions with removable or jump discontinuities are still
integrable, but some functions with vertical asymptotes may not be. You will investigate
various examples of discontinuous functions in Exercises 19 and 20. In Section 5.6 you will
see the somewhat surprising spectacle of certain functions that have vertical asymptotes
and yet are integrable.
Properties of Definite Integrals
Recall from the previous section that Riemann sums measure signed areas: Areas of regions
above the x-axis are counted positively, and areas of regions below the x-axis are counted
negatively. Because definite integrals are defined to be limits of Riemann sums, this means
that definite integrals also automatically count signed areas:
THEOREM 4.10
Definite Integrals Measure Signed Areas
Definite integrals count areas above the x-axis positively and areas below the x-axis
negatively. Algebraically, this means that
b
(a) If f (x) ≥ 0 on all of an interval [a, b], then a f (x) dx ≥ 0.
d
(b) If f (x) ≤ 0 on all of an interval [c, d], then c f (x) dx ≤ 0.
5
For example, if f is the function graphed next, then the definite integral 0 f (x) dx will count
the areas of the regions marked A and C positively and count the area of the region marked
B negatively:
5
0
f (x) dx measures the signed area A − B + C
y
10
5
ⴙ
ⴙ
1
5
2
3
ⴚ
4
5
x
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
345
Definite Integrals
The fact that definite integrals are defined in terms of sums and limits allows us to state
many nice properties of definite integrals. In particular, because definite integrals are just
limits of sums, they behave well with sums and constant multiples, as follows:
THEOREM 4.11
Sum and Constant-Multiple Rules for Definite Integrals
For any functions f and g that are integrable on [a, b] and any real number k,
b
b
b
(a)
( f (x) + g(x)) dx =
f (x) dx +
g(x) dx.
a
b
(b)
a
kf (x) dx = k
a
a
b
f (x) dx.
a
Before we present a formal proof of this theorem, let’s think about the first part of it graphically. The first graph that follows shows a left sum for the area between the graph of
f (x) = x 2 and the x-axis on [0, 2]. The second graph shows a left sum (with the same
n) for the area between g(x) = x and the x-axis on the same interval. In the third graph we
see that the sum of these two left sums is itself a left sum for the graph of f (x)+g(x) = x 2 + x
on [0, 2]. As n approaches infinity, the area of the beige rectangles will approach the area
under f , the area of the blue rectangles will approach the area under g, and the sum of the
areas of the beige and blue rectangles will approach the area under f + g.
Left sum for
f (x) = x 2
Left sum for
f (x) + g(x) = x 2 + x
Left sum for
g(x) = x
y
y
y
6
6
6
5
5
5
4
4
4
3
3
3
2
2
2
1
1
1
1
x
2
1
2
x
1
2
x
The key to the algebraic proof of Theorem 4.11 is that sums and constant multiples
commute with limits and with Riemann sums. Since the definite integral is a limit of Riemann
sums, it must be that sums and constant multiples commute with definite integrals.
Proof. Given a positive integer n, define x =
b−a
and x k = a + kx, and let x k∗ be any point
n
in the subinterval [x k−1 , x k ]. Then by the definition of the definite integral, we have
b
n
( f (x) + g(x)) dx = lim
( f (x k∗ ) + g(x k∗ )) x
← definite integral definition
n→∞
a
k=1
n
( f (x k∗ ) x + g(x k∗ ) x)
= lim
n→∞
= lim
n→∞
= lim
n→∞
=
a
b
k=1
n
f (x k∗ ) x +
k=1
n
n
← split into two sums
k=1
f (x k∗ ) x + lim
k=1
g(x k∗ ) x
← algebra
n→∞
f (x) dx +
b
g(x) dx.
n
g(x k∗ ) x
← sum rule for limits
k=1
← definite integral definition
a
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
346
Chapter 4
November 21, 2012
Definite Integrals
The proof of part (b) is similar: Write the definite integral of y = kf (x) as a limit of Riemann sums,
use properties of sums and limits to pull out the constant k, and then use the definition of the
definite integral again to get k times the definite integral of f . You will write out the details in
Exercise 55.
The following theorem describes what happens if we take a definite integral over a
zero-length interval, or from right to left, or in two pieces:
THEOREM 4.12
Properties Concerning Limits of Integration
If a < b are real numbers and f is integrable on [a, b], then
a
b
a
f (x) dx = 0
(b)
f (x) dx = −
f (x) dx
(a)
a
b
a
Moreover, if c is any real number in [a, b], then
c
b
b
f (x) dx =
f (x) dx +
f (x) dx
(c)
a
a
c
Again, the properties stated in this theorem follow directly from the definition of the definite integral as a limit of Riemann sums. The first property makes sense graphically because
the area under the graph of a function f from x = a to x = a would have a width of zero.
You will prove this property algebraically in Exercise 56. The second property enables us
1
to consider definite integrals like 4 x 2 dx, in which the starting x-value is greater than
the ending x-value. If we integrate “backwards,” say, from x = 4 to x = 1, then we count
the area given by the definite integral negatively. This property can also be proved from the
definition of the definite integral, where in this situation we would have a negative change
a−b
x =
. You will prove the property in Exercise 57.
n
The third part of Theorem 4.12 can also be proved by using the definition of the definite
integral, but instead we present a “convincing argument by picture” in the three figures
shown next. In this picture we have a = 0, b = 3, and c = 1. The area from x = 0 to x = 3
is clearly the sum of the area from x = 0 to x = 1 and the area from x = 1 to x = 3.
Area from x = 0 to x = 1
Area from x = 1 to x = 3
y
Area from x = 0 to x = 3
y
y
5
5
5
4
4
4
3
3
3
2
2
2
1
1
1
2
3
x
1
1
2
3
x
1
2
3
x
Formulas for Three Simple Definite Integrals
In Section 4.5 we will learn a method for quickly calculating many definite integrals. In
the meantime, it will be useful to have a small set of integration formulas that we can use
to solve simple problems. The next theorem describes definite integrals of three common
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
347
Definite Integrals
functions: constant functions, the identity function, and the squaring function. Using these
formulas, together with the sum and constant-multiple rules for definite integrals, we will
be able to quickly evaluate definite integrals for any quadradic functions.
THEOREM 4.13
Definite Integral Formulas
For any real numbers a, b, and c,
b
b
1
c dx = c(b − a)
(b)
x dx = (b2 − a2 )
(a)
a
2
a
(c)
b
1
3
x 2 dx = (b3 − a3 )
a
These formulas are just the tip of the iceberg; in Section 4.5 we will show that we can solve
many definite integrals by using antiderivatives, and in Chapter 5 we will develop methods
for quickly evaluating many more types of definite integrals.
TECHNICAL POINT The definite integral formulas in Theorem 4.13 are true even when the
definite integral in question turns out to be a negative number. For example,
1
1
1
3
x dx = (12 − (−2)2 ) = (1 − 4) = − .
2
−2
2
2
In Section 4.6 we will discuss what it means when a definite integral on an interval [a, b] is
negative.
Proof. We will use the definition of the definite integral to prove part (a) of the theorem. The
proofs of parts (b) and (c) are similar and are left to Exercises 60 and 61. The first two parts of the
theorem can also be argued geometrically; see Exercises 58 and 59. Suppose f (x) = c is a constant
function on [a, b]. Then by the definition of the definite integral, we have
b
a
f (x) dx = lim
n→∞
= lim
n→∞
n
← definition of definite integral
k=1
n
k=1
= lim c
n→∞
= lim c
n→∞
f (x k∗ ) x
c
b−a
n
b−a
n
b−a
n
n
a
← f (x) = c and x = b −
n
1
(n)
= lim c(b − a) = c(b − a)
n→∞
← pull constants out of sum
k=1
← sum formula
← limit of a constant
Examples and Explorations
EXAMPLE 1
Using a sum formula to approximate an area with a large number of rectangles
4
Find the n-rectangle right-sum approximation for 2 (x 2 +1) dx, and then use it to calculate
approximations with n = 100 rectangles and with n = 1000 rectangles.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
348
Chapter 4
November 21, 2012
Definite Integrals
SOLUTION
We’ll find a formula for the n-rectangle approximation and then evaluate at n = 100 and
b−a
2
n = 1000 at the end. Since a = 2 and b = 4, we have x =
= , and therefore
x k = a + kx = 2 +
n
f (x k∗ )x =
k=1
2k
.
n
For the right sum with x k∗ = x k , we have
n
2+
k=1
=
2k 2
+1
n
n
n
2
n
n
16k 10
8k 2
+
+
n3
n2
n
← simple algebra
k=1
=
n
n
n
10 8 2 16 k
+
k
+
1
n3
n2
n
k=1
8
= 3
n
=
k=1
← separate the sums
k=1
16
n(n + 1)(2n + 1)
+ 2
6
n
10
n(n + 1)
+ (n) ← sum formulas
2
n
8n(n + 1)(2n + 1) 16n(n + 1)
+
+ 10.
6n3
2n2
← simplify
Notice that our expression for the n-rectangle right sum no longer has any sigma notation
in it. This means that it is now a simple matter to evaluate the expression at n = 100 and
at n = 1000. At n = 100 we have
800(101)(201) 1600(101)
+
+ 10 = 20.7868,
6, 000, 000
20, 000
and at n = 1000 we have
8000(1001)(2001) 16000(1001)
+
+ 10 = 20.6787.
6, 000, 000, 000
2, 000, 000
CHECKING
THE ANSWER
After all that work it is good to do a reality check. The figure that follows shows a graphing
calculator plot of the area that we just approximated. The area we are considering should
be just a little less than the area of the trapezoid with base from (2, 0) to (4, 0) and top
side connecting (2, 5) and (4, 17). This trapezoid has area
5 + 17
(2)
2
= 21 square units, just
slightly more than our earlier approximations.
Area under f (x) = x2 + 1 on [2, 4]
25
5
0
0
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
EXAMPLE 2
Definite Integrals
Calculating a definite integral exactly
Take the limit as n → ∞ of an n-rectangle right sum to calculate
4
2
349
(x 2 + 1) dx exactly.
SOLUTION
We already did the hard work in the previous example, and all that remains is to take the
limit as n → ∞. The exact area between the graph of f (x) = x 2 +1 and the x-axis on [2, 4] is
4
n
(x 2 + 1) dx = lim
f (x k∗ )x
← definite integral definition
n→∞
2
= lim
n→∞
=
k=1
8n(n + 1)(2n + 1) 16n(n + 1)
+
+
10
← Example 1
6n3
2n2
16
16
62
+
+ 10 = .
6
2
3
← ratios of leading coefficients
In the step where we evaluated the limit, we used the fact that the limit of a “balanced”
rational function with the same degree in the numerator and denominator is equal to the
ratio of leading coefficients.
Notice that we would not have been able to calculate this limit if we did not have a
formula for the Riemann sum from Example 1 that was expressed in terms of n, but with no
“”. Note also that our previous approximation with n = 1000 was quite accurate; it was
62
only approximately 0.012 square unit larger than the actual area,
≈ 20.6667.
3
EXAMPLE 3
Using definite integral formulas
Use properties of definite integrals and the definite integral formulas in Theorem 4.13 to
4
verify the calculation of 2 (x 2 + 1) dx from the previous example.
SOLUTION
Using the sum rule for definite integrals from Theorem 4.11, we can indeed calculate the
same answer that we did in Example 2:
4
4
4
(x 2 + 1) dx =
x 2 dx +
1 dx
← sum rule for definite integrals
2
2
2
1
3
= (43 − 23 ) + 1(4 − 2) =
EXAMPLE 4
62
.
3
← definite integral formulas
Using the algebraic properties of definite integrals
3
3
5
Given that 1 f (x) dx = 4 and 5 2f (x) dx = −3, find 1 f (x) dx. Identify the theorems or
properties that allow each of your steps.
SOLUTION
We pick apart the definite integral we are seeking until we can express it in terms of the
definite integrals that we are given:
5
3
5
f (x) dx =
f (x) dx +
f (x) dx
← Theorem 4.12(c)
1
1
3
=
3
f (x) dx −
1
← Theorem 4.12(b)
f (x) dx
5
3
=
3
f (x) dx −
1
1
2
= 4 − (−3) =
3
1
2f (x) dx
2 5
← Theorem 4.11(b)
11
.
2
← using what was given
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
350
Chapter 4
EXAMPLE 5
November 21, 2012
Definite Integrals
Interpreting definite integrals as signed areas
Use a graph and properties of definite integrals to argue that
5
find a number a so that a (2 − x) dx is exactly zero.
5
1
(2 − x) dx is negative. Then
SOLUTION
The region between the graph of f (x) = 2 − x and the x-axis on [1, 5] consists of one small
triangular region above the x-axis and one larger triangular region below the x-axis:
The region between f (x) = 2 − x and the x-axis on [1, 5]
y
4
3
2
1
3 2 1
1
1
2
3
4
5
6
x
2
3
4
The definite integral will count the region above the x-axis positively and the region below
the x-axis negatively; that is,
2
(2 − x) dx > 0
5
and
1
(2 − x) dx < 0.
2
Since the region above the x-axis is smaller than the region below the x-axis, we must have
5
1
2
(2 − x) dx =
5
(2 − x) dx +
1
(2 − x) dx < 0.
2
5
We will have a (2 − x) dx equal to zero when a = −1, because this is the value for which
the triangle above the x-axis on the interval [a, 1] is the same size as the triangle below
the x-axis on the interval [2, 5]. We could compute the areas of the triangles in question
exactly from the formula for the area of a triangle, but it is sufficient in this case to see the
graph:
The region between f (x) = 2 − x and the x-axis on [−1, 5]
y
4
3
2
1
3 2 1
1
1
2
3
4
5
6
x
2
3
4
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
TEST YOUR
? UNDERSTANDING
351
Definite Integrals
How is a definite integral different than a Riemann sum?
What do definite integrals have to do with areas under curves?
What types of functions are integrable?
b
Why does it make sense that if a < c < b, then
a
f (x) dx =
c
a
f (x) dx +
b
c
f (x) dx?
Why is it not surprising that definite integrals behave well with respect to sums and
constant multiples?
EXERCISES 4.3
Thinking Back
Commuting with sums: What does it mean to say that
derivatives commute with sums? That limits commute
with sums? That sums written in sigma notation commute with sums? Express your answers in words and
algebraically.
Commuting with constant multiples: What does it mean
to say that derivatives commute with constant multiples? That limits commute with constant multiples?
That sums written in sigma notation commute with
constant multiples? Express your answers in words
and algebraically.
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: The left-sum and right-sum approximations are the same if the number n of rectangles is very
large.
5
(b) True or False: −2 (x + 2)3 dx is a real number.
(c) True or False: (5x 2 − 3x + 2) dx is exactly 26.167.
3
−2
(d) True or False: −3 f (x) dx = − 2 f (x) dx.
2
2
(e) True or False: If 0 f (x) dx = 3 and 0 g(x) dx = 2, then
2
0 f ( g(x)) = 6.
2
2
(f) True or False: If 0 f (x) dx = 3 and 0 g(x) dx = 2, then
2
0 f (x)g(x) dx = 6.
1
0
(g) True or False: If 0 f (x) dx = 3 and −1 f (x) dx = 4, then
1
−1 f (x) dx = 7.
4
2
(h) True or False: If 0 f (x) dx = 3 and 2 g(x) dx = 4, then
4
0 ( f (x) + g(x)) dx = 7.
2. Examples: Construct examples of the thing(s) described in
the following. Try to find examples that are different than
any in the reading.
(a) A function that is not integrable on [2, 10].
(b) A function f for which we currently know how to cal2
culate −2 f (x) dx exactly.
(c) A function f for which we currently do not know how
2
to calculate −2 f (x) dx exactly.
3. Fill in the blanks: The signed area between the graph
of a continuous function f and the x-axis on [a, b] is
and is called
represented by the notation
.
the
4. Explain why it makes sense that every Riemann sum for a
continuous function f on an interval [a, b] approaches the
same number as the number n of rectangles approaches
infinity. Illustrate your argument with graphs.
5. Fill in the blanks: The definite integral of an integrable
function f from x = a to x = b is defined to be
b
f (x) dx = lim
,
a
where x =
, and x k∗ is
, xk =
.
6. Explain geometrically what the definition of the definite
integral as a limit of Riemann sums represents. Include a
labeled picture of a Riemann sum (for a particular n) that
illustrates the roles of n, x, x k , x k∗ , and f (x k∗ ). What happens in the picture as n → ∞?
a
7. If f (x) is defined at x = a, then a f (x) dx = 0. Explain why
this makes sense in terms of area.
8. Draw pictures illustrating the fact that if a ≤ c ≤ b, then
c
b
b
f (x) dx +
f (x) dx =
f (x) dx.
a
c
a
Use graphs to determine whether each of the following
definite integrals is equal to a positive number, a negative
number, or zero:
3
3
(x 2 − 4) dx
10.
|x 2 − 4| dx
9.
−3
2π
cos x dx
11.
0
12.
−3
3π /4
−3π /4
CONFIRMING PAGES
cos x dx
18:50
TKmaster2010
WHF00153/FREE087-Taalman
352
Chapter 4
November 21, 2012
Definite Integrals
13. Consider the function f graphed here. Shade in the
regions between f and the x-axis on (a) [−2, 6] and
(b) [−4, 2]. Are the signed areas on these intervals
positive or negative, and why?
y
8
4
4 2
2
4
6
8
x
4
8
14. In a right sum with n = 4 rectangles for the area between the graph of f (x) = x 2 − 1 and the x-axis on [0, 2],
how many rectangles have negative heights? What if
n = 7? In each case, is the estimate for the signed area
positive or negative? Is the exact value of the signed
area positive or negative?
3
(b) Write down an n-rectangle left sum for 0 x 2 dx, and
use algebra and a sum formula to write this sum as a
formula in terms of n.
(c) Use your answers to (a) and (b) to show that the right
3
sum and the left sum for 0 x 2 dx are different for
n = 100 and n = 1000.
(d) Use your answers to (a) and (b) to show that the right
3
sum and the left sum for 0 x 2 dx approach the same
quantity as n → ∞. What does this quantity represent geometrically?
18. The definite integral of a function f on an interval [a, b] is
defined as a limit of Riemann sums. How can it be that
the sum of the areas of infinitely many rectangles that
are each “infinitely thin” is a finite number? On the one
hand, shouldn’t it be infinite, since we are adding up infinitely many rectangles? On the other hand, shouldn’t it
always be zero, since the width of each of the rectangles
is approaching zero as n → ∞?
19. Some discontinuous functions are not integrable. For
example, consider the function f (x) =
(a) Sketch a graph of f on [−2, 2]. What kind of discontinuity does f have, and where?
(b) Why might it be reasonable to think that f is not
integrable on [−2, 2] because of this discontinuity?
(c) What do you think happens to a Riemann sum approximation for the area between f and the x-axis on
[−2, 2] as n → ∞?
(d) Although your intuition probably told you that the
15. Although the definite integral of a sum of functions is
equal to the sum of the definite integrals of those functions, the definite integral of a product of functions is not
the product of two definite integrals.
(a) Use mathematical notation to write the preceding
sentence in this form:
=
,
=
but
.
(b) Choose two simple functions f and g so that you can
calculate the definite integrals of f , g, and f + g on
[0, 1], and show that the sum of the first two definite
integrals is equal to the third.
(c) Find two simple functions f and g such that
1
1
0 f (x)g(x) dx is not equal to the product of 0 f (x) dx
1
and 0 g(x) dx. (Hint: Choose f and g so that you can
calculate the definite integrals involved.)
16. Suppose f is an integrable function [a, b] and k is a
real number. Use pictures of Riemann sums to illustrate
that the right sum for the function kf (x) on [a, b] is k
times the value of the right sum (with the same n) for
f on [a, b]. What happens as n → ∞? What does this
b
exercise say about the definite integrals a f (x) dx and
b
a kf (x) dx?
3
17. Consider the definite integral 0 x 2 dx.
3
(a) Write down an n-rectangle right sum for 0 x 2 dx, and
use algebra and a sum formula to write this sum as a
formula in terms of n.
1
on [−2, 2].
x2
area between the graph of f (x) =
1
and the x-axis
x2
on [−2, 2] was likely to be infinite, this is not always
the case for functions with vertical asymptotes. Surprisingly, as we will see in Section 5.6, the function
g(x) =
1
has an asymptote at x = 0 and yet its area
x 2/3
on [−2, 2] is actually finite! Compare the graphs of
f (x) =
1
1
and g(x) = 2/3 , and speculate as to why
x2
x
this might be the case.
20. Some discontinuous functions are still “nice” enough to
be integrable. For example, consider the function
x, if x ≤ 2
f (x) =
x + 3, if x > 2.
(a) Sketch a graph of f on [0, 5]. What kind of discontinuity does f have, and where?
(b) Why is it reasonable that f is integrable on [0, 5]
despite this discontinuity?
(c) What happens to a Riemann sum approximation for
the area between a function f and the x-axis on [0, 5]
as n → ∞?
Skills
Use geometry (i.e., areas of triangles, rectangles, and circles)
to find the exact values of each of the definite integrals in
Exercises 21–28.
2
1
(2 − x) dx
21.
22.
0(4x − 3) dx
0
2
23.
−3
25.
8
1
−1
24 dx
24.
4
−2
1−
x2
dx
26.
r
−r
CONFIRMING PAGES
|3x + 1| dx
r 2 − x 2 dx
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.3
3
27.
If
6
(1 − |x − 1|) dx
0
3
−2
f (x) dx = 4,
28.
6
−2
2
−2
(3 +
3
f (x) dx = 9,
−2
4 − x 2 ) dx
g(x) dx = 2, and
3 g(x) dx = 3, then find the values of each definite integral
in Exercises 29–40. If there is not enough information, explain
why.
6
6
29.
f (x) dx
30.
g(x) dx
31.
33.
6
−2
3
−2
6
35.
( f (x) + g(x)) dx
3
f (x)g(x) dx
34.
( g(x))2 dx
36.
( f (x) + g(x)) dx
38.
6
−2
39.
(2f (x) − g(x)) dx
32.
3
37.
−2
6
3
(2x 2 − 3g(x)) dx
40.
3
−2
3
−2
6
( g(x) + x) dx
x f (x) dx
(4f (x) − 2) dx
−2
−2
−2
x( f (x) + 3)2 dx
353
For each definite integral in Exercises 41–46, (a) find the
general n-rectangle right sum and simplify your answer with
sum formulas. Then (b) use your answer to approximate the
definite integral with n = 100 and n = 1000. Finally, (c) take
the limit as n → ∞ to find the exact value.
3
5
(5 − x) dx
42.
(2x + 1) dx
41.
2
0
1
43.
3
45.
2x 2 dx
44.
(x + 1)2 dx
46.
−3
2
−3
2
−1
2
3
6
Definite Integrals
x 2 dx
(1 − x 2 ) dx
Calculate the exact value of each definite integral in Exercises 47–52 by using properties of definite integrals and the
formulas in Theorem 4.13.
6
4
(x 2 + 1) dx
48.
(3x + 2) dx
47.
2
2
49.
(9 + 10x − x ) dx
2
4
((2x − 3)2 + 5) dx
0
1
3
50.
5
51.
1
52.
0
(x + 1)2 dx
(3(1 − 2x)2 + 4x) dx
6
Applications
53. Suppose that once again you drive in a car for 40 seconds with velocity v(t) = −0.22t 2 + 8.8t feet per second,
as shown in the graph that follows. Suppose also that
your total distance travelled is equal to the area under the
velocity curve on [0, 40].
Velocity of car
v(t) = −0.22t 2 + 8.8t
54. The function for the standard normal distribution is
1
2
f (x) = √ e−x /2
2π
Its graph is that of the bell curve. Probability and statistics books often have tables like the one following, which
lists some approximate areas under the bell curve:
Areas under the bell curve
v
b
88
10
20
30
40
t
(a) What definite integral would you have to compute in
order to find your exact distance travelled over the
40 seconds of your trip?
(b) Find the exact value of that definite integral by taking
a limit of Riemann sums.
1 b −x 2 /2
e
dx
√
2π −b
0.5
0.3829
1
0.6827
1.5
0.8664
2
0.9545
2.5
0.9876
Use the information given in the table, properties of definite integrals, and symmetry to find
2
1.5
1
1
2
2
e−x /2 dx
e−x /2 dx (b) √
(a) √
2π 1.5
2π −0.5
Proofs
55. Use the definition of the definite integral as a limit of
Riemann sums to prove Theorem 4.11(b): For any function f that is integrable on [a, b] and any real number c,
b
b
a c f (x) dx = c a f (x) dx.
56. Use the definition of the definite integral as a limit of
Riemann sums to prove Theorem 4.12(a): For any funca
tion f and real number a, a f (x) dx = 0.
57. Use the definition of the definite integral as a limit of
Riemann sums to prove Theorem 4.12(b): For any
a
function f that is integrable on [a, b], b f (x) dx =
b
− a f (x) dx.
58. Give a geometric argument to prove Theorem 4.13(a):
b
For any real numbers a, b, and c, a c dx = c(b − a). (Hint:
Use a rectangle.)
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
354
Chapter 4
November 21, 2012
Definite Integrals
59. Give a geometric argument to prove Theorem 4.13(b):
b
1
For any real numbers 0 < a < b, a x dx = (b2 − a2 ).
2
(Hint: Use a trapezoid.)
60. Prove Theorem 4.13(b): For any real numbers a and b, we
b
1
have a x dx = (b2 − a2 ). Use the proof of Theo2
rem 4.13(a) as a guide.
62. Prove that
3
1
(3x + 4) dx = 20 in three different ways:
(a) algebraically, by calculating a limit of Riemann sums;
(b) geometrically, by recognizing the region in question
as a trapezoid and calculating its area;
(c) with formulas, by using properties and formulas of
definite integrals.
61. Prove Theorem 4.13(c): For any real numbers a and b,
b 2
1 3
(b − a3 ). Use the proof of Theorem 4.13(a)
a x dx =
3
as a guide.
Thinking Forward
Functions defined by area accumulation: We know that for fixed
real numbers a and b and an integrable function f , the defb
inite integral a f (x) dx is a real number. For different real
values of b, we get (potentially) different values for the
b
integral a f (x) dx.
b
Make a table of the values of the integral 0 2x dx
corresponding to the values −3, −2, −1, 0, 1, 2, and
3 for b. Conjecture a formula for the relationship
between the values of b and the corresponding value
of the integral.
4.4
What is the word that describes the kind of relationship that exists between the values of b and the
corresponding value of the integral?
Now make a table of the values of the inteb
gral 1 2x dx corresponding to the values −3, −2,
−1, 0, 1, 2, and 3 for b. Conjecture a formula for the
relationship between the values of b and the corresponding value of the integral.
What is the relationship between the formula that
b
describes 0 2x dx and the formula that describes
b
1 2x dx?
INDEFINITE INTEGRALS
The definition of the indefinite integral as a notation for antidifferentiation
Formulas for integrating some basic functions
Guessing and checking to solve integrals of combinations of functions
Antiderivatives and Indefinite Integrals
As we have seen throughout the previous chapters, an antiderivative of a function f is a
function F whose derivative is f . For example, an antiderivative of f (x) = 2x is the function
F(x) = x 2 . Another antiderivative of f (x) = 2x is the function G(x) = x 2 + 3. In fact, any
function of the form x 2 + C is an antiderivative of f (x) = 2x, and these are the only possible
antiderivatives of f (x) = 2x. In general, the antiderivatives of a given function are all related
to each other, so we call the set of all antiderivatives of f the family of antiderivatives of f .
As we showed in Theorem 3.7 of Section 3.2, any two antiderivatives of a function must
differ by a constant. For convenience we restate that theorem now:
THEOREM 4.14
Functions with the Same Derivative Differ by a Constant
If F and G are differentiable functions, then F (x) = G (x) if and only if G(x) = F(x) + C
for some constant C.
In the previous section we defined the definite integral of a function f on an interval
[a, b] as a limit of Riemann sums used in calculating the signed area between the graph of
a function f and the x-axis on an interval [a, b]. We will now define a completely different
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.4
Indefinite Integrals
355
object called the indefinite integral of a function f . This new object will be a family of
functions, not a number, but we will see in Section 4.5 that these two types of integrals are
related, which is why we’ll call them by similar names.
The indefinite integral of a function is a notation for expressing the collection of all
possible antiderivatives of that function:
DEFINITION 4.15
The Indefinite Integral of a Function
The indefinite integral of a continuous function f is defined to be the family of antiderivatives
f (x) dx = F(x) + C,
where F is an antiderivative of f , that is, a function for which F = f .
For example, since an antiderivative of f (x) = 2x is F(x) = x 2, it follows that all antiderivatives of f (x) = 2x are of the form x 2 + C, and therefore that 2x dx = x 2 + C. Note that it
2
would
be equally2 accurate to use a different antiderivative, such as G(x) = x + 3, and say
that 2x dx = (x + 3) + C.
CAUTION
Although the notation and terminology used for indefinite integrals in Definition 4.15 is
similar to what we used for definite integrals in Section 4.3, it is important to note that
at this point we have no proof that the two types of integrals are related. When we see
the Fundamental Theorem of Calculus in Section 4.5, we will make the surprising discovery that the area under a curve is in fact related to families of antiderivatives, and
this relationship will justify why we use such similar notation for two different kinds of
objects.
The “dx” in the notation of Definition 4.15 represents the fact that we are antidifferentiating with respect to the variable x. The constant C represents an arbitrary constant.
The
function f inside the integral notation is called the integrand, and when we find f (x) dx,
we say that we are integrating the function f . The indefinite integral of a function will
often be called simply the integral of that function. The continuity hypothesis is important
(see Exercises 18–20), and we will assume throughout this section that we are working with
intervals where our functions are continuous.
Antidifferentiation Formulas
All of the rules that we have developed for differentiating functions can be used to develop
antidifferentiation rules, which in turn will give us formulas for some common indefinite
integrals. For example, the rule for differentiating power functions says that for any constant
d
k, (x k ) = kx k−1 . The rule for antidifferentiating a power function should “undo” this
dx
process and is given in the following theorem:
THEOREM 4.16
Integrals of Power Functions
(a) If k = −1, then
(b)
x k dx =
1
dx = ln |x| + C.
x
1
x k + 1 + C.
k+1
(See Exercise 18 for a technical point.)
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
356
Chapter 4
November 21, 2012
Definite Integrals
Proof. To prove the first part of the theorem, it suffices to show that if k = −1, then
an antiderivative of x k . (Note that if k = −1, then
need only show that the derivative of
1
is not even defined.) In other words, we
k+1
1
x k + 1 is x k . This is a simple application of the power and
k+1
constant-multiple rules of differentiation:
d
dx
1
x k + 1 is
k+1
1
x k+1
k+1
1
(k + 1)x k = x k .
k+1
=
The second integration formula in the theorem covers the case when k = −1, since x −1 =
formula follows immediately from the fact that
d
1
(ln |x|) = .
dx
x
1
. This
x
The next three theorems describe formulas for antidifferentiating–and thus integrating–
other common types of functions. Each of these formulas can be proved by differentiating;
you will do so in Exercises 73–75.
THEOREM 4.17
Integrals of Exponential Functions
1
(a) If k = 0, then e k x dx = e k x + C.
k
1 x
b + C.
(b) If b > 0 and b = 1, then b x dx =
ln b
2 x dx =
1 x
2
ln2
d
dx
1 x
2
ln2
=
1
(ln 2)2 x
ln2
= 2 x . Notice that both of
the rules in Theorem 4.17 imply that the integral of e x is itself, that is, that e x dx = e x + C.
For example,
because
THEOREM 4.18
Integrals of Certain Trigonometric Expressions
(a)
sin x dx = −cos x + C
(b)
cos x dx = sin x + C
(d)
csc2 x dx = −cot x + C
(c)
sec2 x dx = tan x + C
(e)
sec x tan x dx = sec x + C
(f)
csc x cot x dx = −csc x + C
THEOREM 4.19
Integrals Whose Solutions Are Inverse Trigonometric Functions
1
(a)
dx = sin−1 x + C.
√
1 − x2
1
dx = tan−1 x + C.
(b)
1 + x2
1
(c)
dx = sec−1 x + C.
√
|x| x 2 − 1
Finally, if you are covering hyperbolic and inverse hyperbolic functions in your course
(see the last two subsections of Section 2.6), then you should also be familiar with the
next two sets of antidifferentiation formulas, which follow directly from the differentiation
formulas in Theorems 2.20 and 2.21.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
December 11, 2012
4.4
THEOREM 4.20
Indefinite Integrals
357
Integrals of Hyperbolic Functions*
(a)
sinh x dx = cosh x + C
cosh x dx = sinh x + C
(b)
sech2 x dx = tanh x + C
(c)
THEOREM 4.21
Integrals Whose Solutions Are Inverse Hyperbolic Functions*
1
(a)
dx = sinh−1 x + C
√
2
x +1
1
(b)
dx = cosh−1 x + C
√
x2 − 1
1
dx = tanh−1 x + C
(c)
1 − x2
Antidifferentiating Combinations of Functions
We have rules for differentiating constant multiples, sums, products, quotients, and compositions of functions. Only the constant-multiple rule and the sum rule translate directly
into antidifferentiation rules:
THEOREM 4.22
Constant Multiple and Sum Rules for Indefinite Integrals
(a)
kf (x) dx = k f (x) dx.
(b)
( f (x) + g(x)) dx =
f (x) dx +
g(x) dx.
Proof. Suppose F is any antiderivative of f , that is, F (x) = f (x). Then, by the constant-multiple
rule, kF (x) = kf (x) for any constant k. Furthermore, for any constant D, F(x) + D is also an antiderivative of f , so (F(x) + D) = f (x). Therefore
C
kf (x) dx = kF (x) + C = k F (x) +
= k(F (x) + D) = k f (x) dx.
k
Note that in the calculation we just did, D =
C
is just a constant.
k
Similarly, if F (x) = f (x) and G (x) = g(x), then, by the sum rule, (F(x) + G(x)) = f (x) + g(x),
and therefore
f (x) dx +
g(x) dx = (F (x) + C1 ) + (G (x) + C2 ) = (F (x) + G (x)) + C = ( f (x) + g(x)) dx,
where C = C1 + C2 .
There are no general product, quotient, and chain rules for antidifferentiation. However, if we think of these differentiation rules “backwards,” then we can say something
about certain types of integrands:
CONFIRMING PAGES
0:18
TKmaster2010
WHF00153/FREE087-Taalman
358
Chapter 4
THEOREM 4.23
November 21, 2012
Definite Integrals
Reversing the Product, Quotient, and Chain Rules
(a)
( f (x) g(x) + f (x) g (x)) dx = f (x) g(x) + C
(b)
(c)
f (x)
f (x) g(x) − f (x) g (x)
dx =
+C
( g(x))2
g(x)
f ( g(x)) g (x) dx = f ( g(x)) + C
Of course, the trouble is recognizing when an integrand is in one of these special forms.
If we are lucky enough to recognize an integrand as the result of a product, quotient, or
chain rule calculation, then we can make an educated guess at the integral and check our
answer by differentiating. Repeated guess-and-check is at this point our best strategy for
integrating combinations of functions. The first and third parts of Theorem 4.23 will form
the basis for the methods of integration by parts and integration by substitution that we
will see in Chapter 5.
In general, antidifferentiation is much more difficult than differentiation. We can differentiate every function that we currently know how to write down; however, at this point
we cannot integrate even very simple functions like ln x or sec x. You should think of integration as a puzzle, not a procedure. Unlike differentiation, where it is always clear which
rules to apply, and in which order, it is not always immediately clear how to find a given
integral. We will learn some more methods for calculating integrals in Chapter 5, but even
then we will not be able to calculate all integrals. In fact, as we will see in Section 4.7, some
functions have no elementary antiderivative, which means that their antiderivatives cannot
even be written down in terms of the functions we now know.
Examples and Explorations
EXAMPLE 1
Identifying antiderivatives
Which of the following are antiderivatives of f (x) = x 4 , and why?
(a) 4x 3
1
5
(b) x 5
(c)
1 5
x −2
5
1
5
(d) (x 5 − 2)
SOLUTION
An antiderivative of f (x) = x 4 is a function whose derivative is x 4 . Choice (a) is clearly
the derivative, not an antiderivative, of f (x) = x 4 . The remaining three choices are all antiderivatives of f (x) = x 4 , since each of those functions has derivative x 4 .
EXAMPLE 2
Using algebra to identify and then integrate power functions
1
Find
dx
and
x 3 dx.
x2
SOLUTION
Solving these integrals is an easy application of Theorem 4.16, once we write the integrands
in the form x k :
1
1
1 −1
1
dx
=
x−2 dx =
x−2+1 =
x + C = − + C;
−2 + 1
−1
x
x2
1
1
2
x 3 dx = x 3/2 dx = 3
x 5 + C.
x 3/2+1 = 5 x 5/2 + C =
5
+
1
2
2
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.4
EXAMPLE 3
359
Indefinite Integrals
Using the sum and constant-multiple rules for indefinite integrals
Find (5x 3 − sin x) dx, and show explicitly how the sum and constant-multiple rules for
indefinite integrals are applied.
SOLUTION
The function f (x) = 5x 3 − sin x is a sum of constant multiples of functions whose integrals
we know, namely, the functions x 3 and sin x. By Theorem 4.22, we have
(5x 3 − sin x) dx =
(5x 3 + (−sin x)) dx
5x 3 dx +
=
=5
=5
← write the difference as a sum
← sum rule
(−sin x) dx
x 3 dx −
← constant-multiple rule
sin x dx
1 4
x − (−cos x) + C
4
← antidifferentiation formulas
5
4
= x 4 + cos x + C.
Notice that we added only one constant C in this calculation. Technically there are two
1
such constants, since the family of antiderivatives for x 3 is x 4 + C1 , and the family of
4
antiderivatives for sin x is −cos x + C2 . The C we are using is really the sum C1 + C2 . There
is no need to write the two constants C1 and C2 separately, since the sum of two arbitrary
constants is also an arbitrary constant.
EXAMPLE 4
Recognizing integrands in product, quotient, and chain rule forms
Solve the following indefinite integrals:
(a) (e x + xe x ) dx
(b) 5(x 2 + 1)4 (2x) dx
(c)
2x tan x − x 2 sec2 x
dx
tan2 x
SOLUTION
(a) This integrand looks like it could be the result of a product rule calculation where x and
d
e x are the factors in the product. We check by differentiating, and indeed (xe x ) =
dx
x
x
x
x
x
1e + xe . Therefore (e + xe ) dx = xe + C.
(b) The integrand 5(x2 + 1)4 (2x) looks like it could be the result of a chain rule calculation,
with outside function x 5 and inside function x2 + 1. We can check whether this is the
d
case by differentiating to find ((x 2 + 1)5 ) = 5(x 2 + 1)4 (2x), which is what we had
dx
hoped for. Therefore 5(x 2 + 1)4 (2x) dx = (x 2 + 1)5 + C.
(c) Finally, this integrand could be the result of differentiating a quotient
f (x)
g(x)
where the
denominator is g(x) = tan x. Looking at the numerator of the integrand, we might
guess that f (x) = x 2 . We now check:
wanted for the integrand. Therefore
d
x2
(2x)tanx − x 2 sec2 x
=
, which
dx tanx
(tanx)2
2
2
2
2xtanx − x sec x
x
dx =
+ C.
tanx
tan2 x
CONFIRMING PAGES
is what we
18:50
TKmaster2010
WHF00153/FREE087-Taalman
360
Chapter 4
EXAMPLE 5
November 21, 2012
Definite Integrals
Integrating by educated guess-and-check
Solve the following indefinite integrals:
2
(a)
sin 3x dx
(b)
xe x +1 dx
(c)
1
dx
1 + 4x 2
SOLUTION
(a) At first glance the function sin(3x) does not appear to be of the form f ( g(x))g (x).
Specifically, if we choose f (x) = sin x, so that the inside function is g(x) = 3x,
then the derivative g (x) = 3 does not appear in the integrand. However, since an
antiderivative of sin x is −cos x, we might try F(x) = −cos 3x as a first guess for an
d
antiderivative of sin 3x. Testing this guess, we find that (−cos 3x) = −(−sin 3x)(3) =
dx
3 sin 3x. This is almost the integrand we are looking for, but with an extra constant mul1
1
tiple of 3. Accordingly, we update our guess to F(x) = − cos 3x, so that the at the
3
3
front will cancel out the factor of 3 that appears when we apply the chain rule. We now
have
d
1
1
− cos 3x = − (−sin 3x)(3) = sin 3x,
dx
3
3
and therefore
1
3
sin 3x dx = − cos 3x + C.
(b) The first thing to recognize about this problem is that the function x 2 + 1 is an inside
function in the integrand while its derivative 2x is almost part of the integrand: We
have an x instead of a 2x, so we’re missing only a constant multiple. Since the function
2
e x is its own antiderivative, a good first guess for an antiderivative of xe x +1 might be
2
2
2
d
F(x) = e x +1 . Checking this guess, we find that (e x +1 ) = e x +1 (2x), which is almost
dx
what we want. To get rid of the extra 2, we update our guess to F(x) =
this guess we have
d 1 x 2 +1
e
dx 2
It then follows that
xe x
2
+1 dx
1
2
= ex
1
2
= ex
2
+1
2
+1
(2x) = xe x
2
+1
1 x 2 +1
e
;
2
with
.
+ C. It is important to note that the method
used in this example, whereby we update our guess by dividing by what is “missing,”
works only if what is missing is a constant.
(c) In this example we are searching for a function whose derivative is
1
.
1 + 4x 2
Our inte-
grand is similar in structure to the second integrand in Theorem 4.19, so we will start
the guessing process with tan−1 (x). We guess, check, and re-guess until we get the
derivative we seek:
1
d
(tan−1 x) =
;
dx
1 + x2
1
4
d
(tan−1 (4x)) =
(4) =
;
dx
1 + (4x)2
1 + 16x 2
1
2
d
(tan−1 (2x)) =
(2) =
;
dx
1 + (2x)2
1 + 4x 2
d 1
tan−1 (2x) =
dx 2
Therefore
1
1 + 4x 2
dx =
1
2
1
2
1
1
(2) =
.
1 + (2x)2
1 + 4x 2
tan−1 (2x) + C.
CONFIRMING PAGES
18:50
TKmaster2010
WHF00153/FREE087-Taalman
November 21, 2012
4.4
Indefinite Integrals
361
How is an indefinite integral related to antiderivatives? In what way is a definite integral
TEST YOUR
? UNDERSTANDING
different from an antiderivative?
Which of the six trigonometric functions do we know how to integrate at this point,
and why?
Is the integral of a difference of two functions always equal to the difference of the
integrals of those two functions? Why or why not?
Is the integral of a product of two functions equal to the product of the integrals of
those two functions? Why or why not?
How can we solve integrals of combinations of functions by guessing and checking?
Why is this a valid method?
EXERCISES 4.4
Thinking Back
Antiderivatives: If g (x) = h(x), then is g an antiderivative of h or is h an antiderivative of g?
Definite integrals: State the definition of the definite integral of a function f on an interval [a, b].
The Mean Value Theorem: State the Mean Value
Theorem.
Derivatives and antiderivatives: Find the derivative and an antiderivative of each of the following functions:
√
f (x) = π 2
f (x) = ( 11 x )−2
f (x) =
1
5x
f (x) = sec2 (3x + 1)
Concepts
0. Problem Zero: Read the section and make your own summary of the material.
1. True/False: Determine whether each of the statements that
follow is true or false. If a statement is true, explain why.
If a statement is false, provide a counterexample.
(a) True or False: If f (x) − g(x) = 2 for all x, then f and g
have the same derivative.
(b) True or False: If f (x) − g(x) = 2 for all x, then f and g
have the same ant